Total Ankle Arthroplasty: A Radiographic Outcome Study

Transcription

1 Musculoskeletal Imaging Original Research Lee et al. Outcome of Total nkle rthroplasty Musculoskeletal Imaging Original Research mie Y. Lee 1 lice S. Ha 1 Jonelle M. Petscavage 2 Felix S. Chew 1 Lee Y, Ha S, Petscavage JM, Chew FS Keywords: ankle, arthroplasty, hardware DOI: /JR Received July 18, 2012; accepted after revision September 5, Presented as a scientific talk at the 2012 annual meeting of the RRS and as a scientific poster at the 2012 annual meeting of the Society of Skeletal Radiology. 1 Department of Radiology, University of Washington, ox , 4245 Roosevelt Way NE, Seattle, W ddress correspondence to. S. Ha (aha1@uw.edu). 2 Department of Radiology, Pennsylvania State Hershey Medical Center, Hershey, P JR 2013; 200: X/13/ merican Roentgen Ray Society Total nkle rthroplasty: Radiographic Outcome Study OJECTIVE. Total ankle arthroplasty (T) is becoming a popular alternative to arthrodesis for patients with end-stage ankle arthrosis. Prior outcome studies have primarily focused on surgical findings. Our purpose is to determine the radiographic outcome of T and to correlate it with clinical outcome. MTERILS ND METHODS. In a 9-year retrospective review of patients with T, all available ankle radiographs and clinical data were reviewed. Data analysis included descriptive statistics, Fisher exact test, and Kaplan-Meier survival curves for radiographic and clinical survival. RESULTS. Two hundred sixty-two Ts in 260 patients were reviewed; 55% were in women, and the mean patient age was 61.5 years. The mean radiographic follow-up was 142 weeks, with an average of six radiographs per patient. One hundred sixty-three patients (62.2% of Ts) developed one or more radiographic complications, including periprosthetic lucency (34.0%), hardware subsidence (24.4%), perihardware fracture (11.1%), syndesmotic screw loosening (10.3%), and screw fracture (6.5%). One hundred nineteen patients (45.4% of Ts) developed symptoms of ankle pain or instability, and 71 patients (27.1% of Ts) underwent at least one reoperation. In patients with radiographic complications, 41.7% developed ankle pain and 5.1% developed ankle instability, compared with 23.7% and 2.2%, respectively, of patients with no radiologic complications (p < 0.05). Of the patients with radiographic complications, 33.1% had at least one additional surgery compared with 17.1% of patients without radiographic complications (p < 0.05). CONCLUSION. Radiographic complications are common in patients after total ankle arthroplasty. There is a strong positive association between postoperative radiographic findings and clinical outcome. Knowledge of common postoperative radiographic findings is important for the practicing radiologist. nkle arthrodesis has traditionally been considered the standard surgical treatment of symptomatic endstage ankle arthritis. Total ankle arthroplasty (T) was first introduced in the 1970s as an alternative to ankle fusion. The first-generation ankle implants of the 1970s and 1980s had poor clinical outcomes and thus were largely abandoned [1, 2]. Newer-generation ankle implants attempt to address the biomechanical flaws of the older models. These implants have shown promising initial results [3 6] and have emerged as an increasingly popular treatment of ankle arthritis. Their presumed benefits over arthrodesis include preservation of movement, improvement of gait, and preservation of adjacent joints [2, 7]. Conversion to T is also a via- ble treatment option for patients with painful failed ankle fusions [8]. estic and colleagues [9, 10] have provided two excellent reviews illustrating the postoperative imaging features of T devices. To our knowledge, there is, however, no clinical study in the radiologic literature correlating radiographic findings with clinical outcome. Studies in the surgical literature have primarily focused on surgical and symptomatic outcomes, and only a handful have addressed radiographic evaluations [11 13]. Overall dedicated systematic evaluation of the radiographic outcomes of Ts is lacking. The aim of this study was to determine the radiographic outcome of Ts and to correlate this with clinical outcome JR:200, June 2013

2 Materials and Methods n institutional review board approved retrospective study was performed over a nine-year period at Hershey Medical Center (Pennsylvania State University, Hershey, P) and over a 5-year period at Harborview Medical Center (University of Washington, Seattle, W). Patients were found by use of a PCS (ISite, Philips Healthcare) with a search tool (Primordial, Primordial Design) at Pennsylvania State. database of patients who underwent ankle arthroplasties was provided by the orthopedic surgery department at the University of Washington. Data were collected from a review of the patient s electronic medical records, including preoperative clinical notes, operative reports, and follow-up clinical notes. Demographic data collected included patient sex, age, and indication for arthroplasty. ll available preoperative, intraoperative, and postoperative ankle radiographs were reviewed by a fellowship-trained attending musculoskeletal radiologist. Specific postoperative radiographic findings recorded included periprosthetic lucency of at least 2 mm, hardware subsidence of at least 4 mm, periprosthetic fracture, hardware fracture, heterotopic ossification, increased valgus or varus angulation of the ankle, and ankle gutter narrowing. Lucency adjacent to the lateral aspect of the tibial tray in patients with syndesmotic fusion was considered as expected mechanical lysis [11], unless radiolucency increased after mature fusion across the syndesmosis. Radiographic findings were correlated with the clinical information obtained from the postoperative clinical notes. Patient s symptoms, primarily ankle pain or a sensation of ankle instability or imbalance, were recorded. Reports from all subsequent ankle reoperations were reviewed. Statistical significance was determined with Fisher exact test. Kaplan-Meir survivorship analysis, with radiographic failure and clinical failure as endpoints, was performed. Results From January 2003 to December 2011, 118 ankle arthroplasties were performed in 116 patients at Pennsylvania State University Medical Center. From January 2005 to February 2010, 149 ankle arthroplasties were performed in 149 patients at Harborview Medical Center, University of Washington. Five ankle arthroplasties performed in five patients were excluded from this study because of the absence of radiographic followup. Thus, a total of 262 ankle arthroplasties in 260 patients were reviewed. The mean patient age at the time of the primary arthroplasty was 61.5 years (range, years). Fifty-five percent were women, and Outcome of Total nkle rthroplasty TLE 1: Indications for Total nkle rthroplasty (T) Indication No. (%) of Ts (n = 262) Posttraumatic osteoarthritis 122 (46.6) Primary osteoarthritis 120 (45.8) Rheumatoid arthritis 12 (4.6) Failed prior ankle fusion 4 (1.5) Charcot foot 1 (0.4) vascular necrosis 1 (0.4) Failed T 1 (0.4) 45% were men. Common indications for arthroplasty (Table 1) included posttraumatic osteoarthritis (46.6%), primary osteoarthritis (45.8%), rheumatoid arthritis (4.6%), and failed prior ankle fusion (1.5%). Fifty-two percent of arthroplasties were performed on the left ankle. There were 197 gility (DePuy), 56 Salto Talaris (Tornier), and nine STR (Small one Innovations) implants (Fig. 1). Some patients had concomitant procedures performed at the time of ankle arthroplasty. The most common of these adjuvant procedures included subtalar fusion (31%); other joint fusion in the foot (20.8%); old hardware removal (30.5%); realignment osteotomies in the calcaneus, tibia, or fibula (13.9%); chilles tendon lengthening (59.7%); gastrocnemius recession (22.9%); and tendon transfers (13.2%). Syndesmotic fusion was performed with all gility implants. The mean radiographic follow-up was 142 weeks (range, weeks), with an average of six follow-up radiographs per patient. One hundred sixty-three patients (62.2% of Ts) developed one or more radiographic complications, with a mean time of 74 weeks to the first radiographic complication. Radiographic complications included lucency around the arthroplasty (34.0%), hardware subsidence (24.4%), perihardware fracture (11.1%), syndesmotic screw loosening (10.3%), hardware fracture (6.5%), heterotopic ossification (5.7%), increased varus or valgus angulation of the ankle (5.4%), ankle gutter narrowing (5.4%), and syndesmotic nonunion (2.7%) (Figs. 2 5 and Table 2). Perihardware lucency around the arthroplasty was by far the most common radiographic complication. Of the 89 patients with lucency around the arthroplasty, 89% had lucency around the tibial component and 11% had lucency around the talar component. One hundred nineteen patients (45.4% of Ts) eventually developed symptoms of ankle pain or instability, with a mean time of 57 weeks to symptom development. Seventy-one patients (27.1% of Ts) required at least one additional surgery with an average of 121 weeks to the first reoperation (Fig. 5). This included 17 ankle débridements or arthrotomies, 13 screw or plate removals, 11 syndesmotic fusion revisions, 19 ankle arthroplasty revisions (defined as polyethylene or metallic component replacement), eight realignment osteotomies, two total ankle joint fusions, six bone grafts for periprosthetic lucencies or cysts, and two abscess irrigations. Overall, complication rates were similar between our two institutions. Of patients with radiographic complications, 41.7% developed ankle pain and 5.1% developed ankle instability, compared with 23.7% and 2.2% of patients with no radiologic complications, respectively (p < 0.05). Of patients with radiographic complications, 33.1% had at least one additional surgery compared with 17.1% of patients without radiographic complications (p < 0.05) (Fig. 6). This statistical significance held true for both institutions and all arthroplasty types. Patients with both hardware subsidence and perihardware lucency together had higher rates of ankle pain or instability (82.5%) compared with patients with subsidence alone (28.6%) or perihardware lucency alone (70%). Similarly, patients with both radiographic complications together had higher rates of ankle reoperation (42.5%) compared with patients with subsidence alone (17.1%) or perihardware lucency alone (39.5%). Kaplan-Meier survival curve analysis was performed for clinical survival and radiographic abnormality (Fig. 7). Patients who were lost to follow-up were used as censored data in the curve creation. Clinical failure of arthroplasty was defined as the time of the second surgery. Radiographic failure was defined as the time to the first radiographic abnormality. t about 90 weeks after initial placement, 50% of T implants had radiographic abnormalities. y 200 weeks, only 20% were deemed to be radiogra- JR:200, June

3 Lee et al. TLE 2: Radiographic Complications fter Total nkle rthroplasty (T), by Implant Model Radiographic bnormality gility (DePuy) (n = 197) Salto Talaris (Tornier) (n = 56) Total (n = 262) Total no. of cases with radiographic abnormality 140 (71) 23 (41) 163 (62.2) Perihardware lucency (> 2 mm) 80 (41) 9 (16) 89 (34.0) Subsidence 55 (28) 9 (16) 64 (24.4) Perihardware fracture 25 (13) 4 (7) 29 (11.1) Syndesmotic screw loosening 27 (14) 27 (10.3) Hardware fracture 17 (9) 0 (0) 17 (6.5) Heterotopic ossification 9 (5) 6 (11) 15 (5.7) Increased varus or valgus ankle 13 (7) 1 (2) 14 (5.4) nkle gutter narrowing 13 (7) 1 (2) 14 (5.4) Syndesmotic nonunion or fracture 7 (4) 7 (2.7) Note Data are no. of abnormalities (percentage of Ts). In total, 262 Ts were studied, and 163 had radiographic abnormalities. Of note, none of the nine STR (Small one Innovations) replacement cases had radiographic complications. Dashes indicate not applicable. phically normal. Overall, on the basis of the Kaplan-Meier survival analysis, radiographic survival was worse than the clinical survival. Discussion T is becoming an increasingly popular treatment of end-stage ankle arthritis. We present a retrospective study radiographically describing the largest number of implants to date. Radiographic complication rates are correlated with clinical outcome for the first time, to our knowledge. Our data show that radiographic complications are common in patients after T (163 patients/262 Ts [62.2%]). The two most common complications are perihardware lucency and hardware subsidence; 45.4% of patients developed adverse clinical symptoms, and 27.1% required reoperation. Radiographic findings had a statistically significant positive association with clinical complications after T, with a relatively high rate of ankle symptoms and reoperation in patients with radiographic complications. Our data further validate radiography as an important tool in the postoperative evaluation of patients who undergo T. Previous studies in the surgical literature have largely focused on surgical and symptomatic outcome, and only a few studies have also evaluated postoperative radiographic findings. One recent study [13] evaluated the radiographic findings of three-component mobile-bearing arthroplasties in western France. lthough these implant types are popular in Europe [7], two-component designs, such as the gility implant, are by far the most commonly used in the United States. Our study included three types of T (gility, Salto Talaris, and STR) but was not intended to compare different implant designs against each other. Each design type s data were separately calculated for statistical significance. The gility implant is unique for its use of syndesmotic screws; therefore, syndesmotic screw perihardware lucency and screw fractures were uniquely seen in this patient group. We found rates of subsidence (24%) to be similar to those reported in the surgical literature, which ranged from 14% to 45% [11 13]. Those prior studies similarly found that the most common radiographic complication was radiolucency around the arthroplasty, with rates ranging from 72% to 86%. This much higher rate of perihardware lucency, compared with that in our study (34%), is likely because most of those studies also included lucencies measuring smaller than 2 mm. We found rates of perihardware lucency and subsidence to be higher than those reported for primary total arthroplasties of other joints, such as the shoulder and hip. For example, perihardware lucency has been reported to occur in 12.5% of rough stem total hip arthroplasties and 2.2% of satin stem hip arthroplasties [14]. One study on shoulder arthroplasties found rates of perihardware lucency and subsidence to be 21% and 0%, respectively [15]. Higher rates of radiographic complication in Ts are likely secondary to the large axial loading forces on the ankle, with relatively few surrounding muscles for additional support. Compressive forces on the ankle are estimated to be three to five times the body weight during normal ambulation [1]. Our study s rate of reoperation (27.1%) after T is similar to the reoperation rate of 28% for ankle arthroplasties performed from 1995 to 2001 at Harborview Medical Center, as previously published by our surgical colleagues [16]. retrospective study comparing T to ankle fusion found a similar rate of revision surgery (23%) for T at 5 years [17]. That study also found a significantly lower rate of major revision surgery (11%) for ankle arthrodesis compared with T. It is worth noting, however, that some investigators have argued that revision surgery in T does not imply failure and that obligatory reoperation is anticipated in the natural evolution of ankle implants [7]. Regardless of the definition of failure, radiographic complications in Ts are strongly correlated with the need for repeat surgery. Limitations of this study include its reliance on retrospective clinician-reported physical examinations and nonstandardized qualitative reporting of patients subjective symptoms. Furthermore, the ankle arthroplasties were performed by multiple orthopedic surgeons at two institutions, likely confounding the results by differences in surgical technique and levels of experience. However, rates of complication were similar at our two institutions. nother limitation of this study includes loss to follow-up of patients at different time points, which required censoring in the survival analysis. Patients in this study were followed almost exclusively by radiographs, with only a handful undergoing CT to evaluate for postoperative complication. No patients underwent postoperative ultrasound or MRI. The utility of these other modalities was thus not evaluated in this study. The number of Ts will likely continue to increase and will undoubtedly be more frequently encountered in radiology practice. We present a large sample size follow-up study on T with emphasis on radiographic findings JR:200, June 2013

4 Outcome of Total nkle rthroplasty Given the strong positive association between radiographic findings and clinical outcome, knowledge of the relevant abnormal radiographic findings is important for the practicing radiologist. References 1. Gill LH. Challenges in total ankle arthroplasty. Foot nkle Int 2004; 25: Gougoulias NE, Khanna, Maffulli N. History and evolution in total ankle arthroplasty. r Med ull 2009; 89: Claridge RJ, Sagherian H. Intermediate term outcome of the gility total ankle arthroplasty. Foot nkle Int 2009; 30: Nunley J, Caputo M, Easley ME, Cook C. Intermediate to long-term outcomes of the STR total ankle replacement: the patient perspective. J one Joint Surg m 2012; 94: onnin M, Gaudot F, Laurent JR, Ellis S, Colombier J, Judet T. The Salto total ankle arthroplasty: survivorship and analysis of failures are 7 to 11 years. Clin Orthop Relat Res 2011; 469: Pyevich MT, Saltzman CL, Callaghan JJ, lvine FG. Total ankle arthroplasty: a unique design two to twelve-year follow-up. J one Joint Surg m 1998; 80: Easley ME, dams S Jr, Hembree WC, DeOrio JK. Results of total ankle arthroplasty. J one Joint Surg m 2011; 93: Hintermann, arg, Knupp M, Valderrabano V. Conversion of painful ankle arthrodesis to total ankle arthroplasty: surgical technique. J one Joint Surg m 2009; 91: estic JM, ancroft LW, Peterson JJ, Kransdorf MJ. Postoperative imaging of the total ankle arthroplasty. Radiol Clin North m 2008; 46: estic JM, Peterson JJ, DeOrio JK, ancroft LW, erquist TH, Kransdorf MJ. Postoperative evaluation of total ankle arthroplasty. JR 2008; 190: Knecht SI, Estin M, Callaghan JJ, et al. The gility total ankle arthroplasty: seven to sixteen-year followup. J one Joint Surg m 2004; 86-: Kopp FJ, Patel MM, Deland JT, O Malley MJ. Total ankle arthroplasty with the gility prosthesis: Fig. 1 Normal appearance of total ankle arthroplasty (T) as seen on anteroposterior radiographs., Patient with gility (DePuy) T implant., Patient with Salto Talaris (Tornier) T implant. C, Patient with STR (Small one Innovations) T implant. clinical and radiographic evaluation. Foot nkle Int 2006; 27: Preyssas P, Toullec E, Henry M, Neron J, Mabit C, rilhault J. Total ankle arthroplasty: threecomponent total ankle arthroplasty in western France a radiographic study. Orthop Traumatol Surg Res 2012; 98(suppl 4):S31 S Della Valle G, Zoppi, Peterson MG, Salvati E. rough surface finish adversely affects the survivorship of a cemented femoral stem. Clin Orthop Relat Res 2005; 436: Sanchez-Sotelo J, O Driscoll SW, Torchia ME, Cofield RH, Rowland CM. Radiographic assessment of cemented humeral components in shoulder arthroplasty. J Shoulder Elbow Surg 2001; 10: Spirt, ssal M, Hansen ST Jr. Complications and failure after total ankle arthroplasty. J one Joint Surg m 2004; 86-: SooHoo NF, Zingmond DS, Ko CY. Comparison of reoperation rates following ankle arthrodesis and total ankle arthroplasty. J one Joint Surg m 2007; 89: C JR:200, June

5 Lee et al. C D Fig year-old man with perihardware lucency after total ankle arthroplasty (T) with gility (DePuy) implant., nteroposterior radiograph of ankle obtained immediately after T shows normal appearance of implant., nteroposterior radiograph obtained 3 years later shows new lucencies along medial and lateral aspects of tibial component. Fig year-old woman with subsidence after total ankle arthroplasty (T) with Salto Talaris (Tornier) implant. and, nteroposterior () and lateral () radiographs obtained immediately after T show normal appearance of implant. C and D, nteroposterior (C) and lateral (D) radiographs obtained 4 years later show subsidence of talar component JR:200, June 2013

6 Outcome of Total nkle rthroplasty Fig year-old man with syndesmotic hardware fracture after total ankle arthroplasty (T) with gility (DePuy) implant., nteroposterior radiograph obtained immediately after T shows normal appearance of implant., nteroposterior radiograph obtained 3 years later shows fracture of syndesmotic screws with associated widening of tibiofibular syndesmosis. Fig year-old woman with perihardware fracture after total ankle arthroplasty (T) with gility (DePuy) implant., nteroposterior radiograph obtained 5 years after T placement shows medial malleolar fracture. Note is also made of perihardware lucency along lateral aspect of tibial component., nteroposterior radiograph obtained 3 months later after plate-and-screw fixation of medial malleolar fracture with lateral cyst curettage and packing. JR:200, June

7 Lee et al. Clinical Complication Rate (%) Radiographic complication No radiographic complication 41.7% Pain 23.7% 5.1% 2.2% Instability 33.1% 17.1% dditional Surgery Fig. 6 Strong association between radiographic abnormalities and clinical symptoms after total ankle arthroplasty. ll three comparisons showed statistical significance (p < 0.05). Surviving Ts (%) Gray = radiographic survival lack = clinical survival Time Since Surgery (wk) Fig. 7 Kaplan-Meier survival analysis curve for total ankle arthroplasties (Ts). Graph shows clinical survivorship (black line) compared with radiographic survivorship (gray line). Clinical survivorship is defined as time to revision or retrieval surgery. Radiographic survivorship is defined as time of first radiographic abnormality JR:200, June 2013