Ankle Arthrodesis: a Retrospective Analysis Comparing Single Column, Locked Anterior Plating to Crossed Lag Screw Technique

The Journal of Foot & Ankle Surgery xxx (2017) 1–4

Contents lists available at ScienceDirect

The Journal of Foot & Ankle Surgery

journal homepage: www.jfas.org

Original Research Ankle Arthrodesis: A Retrospective Analysis Comparing Single Column, Locked Anterior Plating to Crossed Lag Screw Technique

Mark A. Prissel, DPM, AACFAS 1, G. Alex Simpson, DO 2, Sean A. Sutphen, DO 3, Christopher F. Hyer, DPM, MS, FACFAS 4, Gregory C. Berlet, MD 1

1 Attending Physician, Orthopedic Foot and Ankle Center, Westerville, OH 2 Attending Physician, Front Range Orthopaedics, Colorado Springs, CO 3 Resident, OhioHealth-Doctor’s Hospital, Columbus, OH 4 Attending Physician and Fellowship Director, Orthopedic Foot and Ankle Center, Westerville, OH article info abstract

Level of Clinical Evidence: 3 Ankle arthrodesis is performed to eliminate pain due to end-stage osteoarthritis, regardless of etiology. This procedure remains the reference standard treatment for end-stage ankle arthritis, despite recent advance- Keywords: arthritis ments in total ankle replacement. The objective of the present study was to retrospectively evaluate the fusion radiographic and clinical fusion rates and time to bony fusion for patients who underwent ankle arthrodesis nonunion using an anterior approach with a single column locked plate construct versus crossed lag screws. We iden- surgery tified 358 patients who had undergone ankle arthrodesis from January 2003 to June 2013. Of the 358 patients, tibiotalar 83 (23.2%) met the inclusion criteria for the present study. Of the 83 included patients, 47 received locked anterior (or anterolateral) plate fixation, and 36 received crossed lag screw constructs. The overall nonunion rate was 6.0% (n ¼ 5), with 1 nonunion in the anterior plate group (2.1%) and 4 nonunions in the crossed lag screw group (11.1%; p ¼ .217). No differences were identified between the 2 groups for normal talocrural angle [c2 (1) ¼ 0.527; p ¼ .468], normal tibial axis/talar ratio [c2 (1) ¼ 0.004; p ¼ .952], and lateral dorsiflexion angle (p ¼ .565). Based on our findings in similar demographic groups, ankle arthrodesis using locked anterior plate fixation is a safe technique with similar complication rates and radiographic outcomes to those of crossed lag screws. Ó 2017 by the American College of Foot and Ankle Surgeons. All rights reserved.

Ankle arthrodesis is performed to eliminate pain due to end-stage in fusion have been limited in published studies thus far (3,5,7–10). osteoarthritis, regardless of etiology. This procedure remains the Recently, a biomechanical comparison determined a superior bending reference standard treatment for end-stage ankle arthritis, despite stiffness with a locked plate and crossed screw construct compared recent advancements in total ankle replacement. Ankle arthrodesis is with either a locked plate or crossed screw construct alone (11). The also indicated for failed ankle arthroplasty and revision ankle objective of the present study was to retrospectively evaluate the arthrodesis (1–7). Ankle fusion involves the bony consolidation of the radiographic and clinical fusion rates and time to bony fusion for tibiotalar articulation and, in some instances, incorporation of the patients who underwent ankle arthrodesis using an anterior distal fibula into the construct (2,4–8). Several various fixation tech- approach with a single column locked plate construct versus crossed niques and constructs have been historically described. Currently, the lag screws. most common methods of fixation include internal screws, plate fixation, or a combination of the 2. Additionally, anatomic ankle fusion Patients and Methods plates are available that are intended for various surgical approaches. fi After institutional review board approval, a retrospective medical record and The results of screw versus plate xation and different types of plates radiographic review was performed of all operative patients within a single foot and ankle specialty practice (Orthopedic Foot and Ankle Center, Westerville, OH) from January 2003 to June 2013 who had undergone ankle arthrodesis. The patients were Financial Disclosure: None disclosed. identified using the Current Procedural Terminology (American Medical Association, Conflict of Interest: Drs. Berlet and Hyer report that they are consultants to Chicago, IL) code 27870, ankle arthrodesis. The exclusion criteria were the presence of Wright Medical Technology, Inc. Dr. Prissel reports that he is a consultant to NovaStep, diabetes or Charcot neuroarthropathy, a lateral or transfibular approach for ankle Inc. arthrodesis, presentation for revision ankle arthrodesis, a lack of 3 months of follow- Address correspondence to: Christopher F. Hyer, DPM, MS, FACFAS, Orthopedic up data available, the presence of peripheral neuropathy, concomitant adjacent joint Foot and Ankle Center, 300 Polaris Parkway, Suite 2000, Westerville, OH 43082. arthrodesis, fixation constructs other than anterior plating or crossed lag screws, and E-mail address: [email protected] (C.F. Hyer). patient age <18 years. All patients identified for the anterior plate group also had 1 lag

1067-2516/$ - see front matter Ó 2017 by the American College of Foot and Ankle Surgeons. All rights reserved. http://dx.doi.org/10.1053/j.jfas.2017.01.007 2 M.A. Prissel et al. / The Journal of Foot & Ankle Surgery xxx (2017) 1–4 screw placed before application of the anterior locking plate. A radiographic review was When comparing the 2 groups, no statistically significant differ- performed to assess the time in days to union, determined by bony trabeculation about ences were identified in gender [c2 (1) ¼ 0.756; p ¼ .385], age at 3 cortices. Additionally, secondary fusion indicators, including the time in days to full surgery (t ¼ 0.86, df ¼ 81; p ¼ .391), body mass index (t ¼ –0.35, weightbearing and time in days to wearing regular shoe gear were recorded. Radio- graphic measurements were performed at confirmed fusion, including the talocrural df ¼ 81; p ¼ .725), tobacco use (p ¼ .318), workers’ compensation angle, tibial axis/talar ratio, and lateral dorsiflexion angle (Fig.) (12–14). (p > .999), history of rheumatoid arthritis (p > .999), operative side The demographic and clinical characteristics were recorded using the mean [c2 (1) ¼ 0.207; p ¼ .649], biologic augmentation [c2 (1) ¼ 1.396; standard deviation for continuous variables and frequencies and percentages for cat- ¼ c2 ¼ fi p .238], overall incidence of hardware removal [ (1) 0.160; egorical variables. To assess the differences between patients with plate xation and ¼ ¼ > patients with lag screws alone, the c2 and Fisher’s exact test were used to compare p .689], infection (p .971), need for revision (p .999), or categorical variables, and 2-sample t tests were used to compare normally distributed neurosensory disturbance (p ¼ .945). Between the 2 groups, we found continuous variables. The follow-up duration, time to full weightbearing, and time to no statistically significant differences in the median interval to the regular shoe wear between the 2 groups were compared using Wilcoxon rank sum follow-up examination (plate fixation, median 240.0 days, range 92.0 tests. The time to fusion among patients with plate fixation and those with only lag to 1087.0; lag screw, median 239.5 days, range 92.0 to 3534.0; screws were compared using the log-rank test. Statistical significance was defined at 2 the 5% (p .05) level. p ¼ .989), time to fusion [c (1) ¼ 0.600; p ¼ .439), median interval to full weightbearing (plate fixation, median 49.0 days, range 9.0 to 97.0; lag screw, median 48.0 days, range 8.0 to 97.0; p ¼ .606) or interval to Results regular shoe wear (plate fixation, median 97.0 days, range 55.0 to 498.0; lag screw, median 101.0 days, range 62.0 to 238.0; p ¼ .097; We identified 358 patients who had undergone ankle arthrodesis Table 1). from January 2003 to June 2013. Of the 358 patients, 83 (23.2%) met A postoperative radiographic review was performed for radio- the inclusion criteria for present study. Of the 83 included patients, 47 graphic union and radiographic measurements at fusion, including received locked anterior (or anterolateral) plate fixation, and 36 the talocrural angle, tibial axis/talar ratio, and lateral dorsiflexion received crossed lag screw constructs. Of the 36 crossed lag screw angle. No differences were identified between the 2 groups for the constructs, various surgical approaches were used, including anterior normal talocrural angle [c2 (1) ¼ 0.527; p ¼ .468], normal tibial axis/ (n ¼ 14), mini-open (n ¼ 8), arthroscopic (n ¼ 7), medial (n ¼ 6), and talar ratio [c2 (1) ¼ 0.004; p ¼ .952], and lateral dorsiflexion angle posterior (n ¼ 1). Of the 36 crossed lag screw constructs, 18 patients (p ¼ .565; Table 2). had 2 screws (50%) and 18 patients had 3 screws (50%). The overall nonunion rate was 6.0% (n ¼ 5), with 1 nonunion in the anterior plate Discussion group (2.1%) and 4 nonunions in the crossed lag screw group (11.1%; p ¼ .217). Regarding the anterior plate group, hardware removal The purpose of our study was to evaluate the fusion rate, time to occurred in 10 patients (21.3%), including 6 screw removals (6 of 47 bony union, radiographic alignment, and incidence of complications screws; 12.8%), 2 plate removals (2 of 47 plates; 4.3%) and complete after ankle arthrodesis when comparing 2 different fixation constructs. removal of 2 plate and crossed screw constructs (2 of 47; 4.3%). A total Against the current body of published data, our cohort of single column plate removal rate of 8.5% (4 of 47) was identified. For the crossed anterior locked plate fixation represents one of the largest reported screw group, hardware removal occurred in 25% of patients [9 of 36; series. Anticipated concerns about this technique, such as an increased c2 (1) ¼ 4.20; p ¼ .041]. Considering our included sample size for each incidence of neurosensory disturbance or a more frequent requirement group with regard to nonunion, we had only 40.8% power to detect a of hardware removal when using plate fixation, were simply unfounded significant difference, with an a of 0.05, because our exclusionary compared with our crossed lag screw cohort. Also, the incidence of plate criteria were stringent. A post hoc power analysis was performed removal was significantly lower than the rate of hardware removal in assuming the same proportions and a level for nonunion (2.1% versus the crossed lag screw group. Also, the radiographic position of the ankle 11.1%), which concluded 119 patients would be required in each group fusion was similar between the 2 groups. Concern could be postulated to obtain 80% power. for a tendency for anterior translation of the talus using plate fixation to

Fig.. Radiographic images demonstrating the measurements used to conﬁrm fusion: (A) talocrural angle, (B) lateral dorsiﬂexion angle, and (C) tibial axis/talar ratio. M.A. Prissel et al. / The Journal of Foot & Ankle Surgery xxx (2017) 1–4 3

Table 1 Table 2 Patient characteristics Radiographic results and secondary fusion indicators

Characteristic Plate Fixation Lag Screws Alone p Value Variable Overall Plate Lag Screws p Value (n ¼ 47) (n ¼ 36) (n ¼ 78)* Fixation Alone ¼ ¼ Demographic (n 46) (n 32) Male gender 28 (59.6) 18 (50.0) .385 Interval to radiographic fusion .439 Age at surgery (yr) 55.0 13.8 52.7 9.4 .391 6 wk Postoperatively 11 (14.1) 9 (19.6) 2 (6.3) Body mass index (kg/m2) 32.1 7.8 32.7 5.7 .725 12 wk Postoperatively 45 (57.7) 24 (52.2) 21 (65.6) Tobacco use 7 (14.9) 2 (5.6) .318 24 wk Postoperatively 18 (23.1) 11 (23.9) 7 (21.9) Workers’ compensation 5 (10.6) 4 (11.1) >.999 36 wk Postoperatively 4 (5.1) 2 (4.4) 2 (6.3) History of rheumatoid arthritis 3 (6.4) 3 (8.3) >.999 Secondary fusion indicators Primary surgical data Interval to full weightbearing .606 Side .689 (days) Right 31 (66.0) 22 (61.1) Median 48.0 49.0 48.0 Left 16 (34.0) 14 (38.9) Range 8.0 to 97.0 9.0 to 97.0 8.0 to 97.0 Biologic augmentation 34 (72.3) 30 (83.3) .238 Interval to regular shoe wear .097 Year of surgery NA (days) 2003 0 (0.0) 2 (5.6) Median 99.5 97.0 101.0 2004 0 (0.0) 5 (13.9) Range 55.0 to 498.0 55.0 to 498.0 62.0 to 238.0 2005 0 (0.0) 8 (22.2) Radiographic measurements 2006 0 (0.0) 7 (19.4) at fusion 2007 0 (0.0) 4 (11.1) Normal talocrural angle 50 (64.1) 31 (67.4) 19 (59.4) .468 2008 1 (2.1) 4 (11.1) (83 4) 2009 0 (0.0) 3 (8.3) Normal tibial axis/talar ratio 32 (41.0) 19 (41.3) 13 (40.6) .952 2010 6 (12.8) 0 (0.0) (35 3) 2011 9 (19.2) 3 (8.3) Normal lateral dorsiflexion 70 (89.7) 40 (87.0) 30 (93.8) .565 2012 16 (34.0) 0 (0.0) angle (90 5) 2013 15 (31.9) 0 (0.0) Data presented as n (%). Complications * Of 83 patients, 78 (94.0%) experienced fusion; 1 patient with plate fixation and 4 Hardware removal 10 (21.3) 9 (25.0) .689 patients with lag screws alone experienced nonunion. Nonunion 1 (2.1) 4 (11.1) .217 Malunion 0 (0.0) 0 (0.0) NA Infection 4 (8.5) 4 (11.1) .971 Revision 1 (2.1) 1 (2.8) >.999 fixation constructs, reducing the homogeneity of that group. Addi- Neurovascular injury 0 (0.0) 0 (0.0) NA tionally, our selection of Current Procedural Terminology codes to Neurosensory disturbance 4 (8.5) 2 (5.6) .945 evaluate could have been different from an alternate surgeon group Abbreviation: NA, not applicable. attempting to perform a similar study. Moreover, clinical documen- Data presented as n (%) or mean standard deviation. tation and serial postoperative radiographic images were used to assess the status of union in the present study because computed tomography scans were not available for all included patients; thus, it ensure a proper footprint for screw placement for the dedicated talar is possible that the nonunion rate was underreported. Also, the screws, which would increase the midfoot joint stresses secondary to an radiographic measurements we used are intended for assessment of a extended lever arm. However, the tibial axis/talar ratio was similar native ankle joint, because, to the best of our knowledge, specific between the 2 groups. The radiographic measurements we used have measurements for ankle arthrodesis do not exist. been described for assessment of an ankle joint that has not undergone In conclusion, based on our findings from similar demographic fusion; however, to our knowledge, specific radiographic measure- groups, ankle arthrodesis using locked anterior plate fixation is a safe ments for the ankle joint after arthrodesis do not exist currently. The technique with similar complication rates and radiographic outcomes lateral dorsiflexion measurement is used to assess the incidence of a to those of crossed lag screws. Although 1 of the largest cohorts to plantarflexed fusion, which places additional stresses on the midfoot date, the sample size for the present study was relatively small to offer and adjacent joints. The tibial axis/talar ratio resulted in a slightly causal statements regarding the outcomes between these 2 accepted greater than average value in both groups, demonstrating an intentional techniques. Thus, further, appropriately powered, retrospective (or slight posterior translation of the talus relative to the tibia, again pro- ideally prospective) studies are warranted to compare these tech- tecting the midfoot. niques to ultimately determine whether a difference in the incidence Chalayon et al (15) recently reported the outcomes of a large of nonunion can be identified. retrospective series of open ankle arthrodesis using various fixations constructs and approaches. Their series reported on 47 anterior plate constructs, a patient volume identical to that in the present study and References for both studies, the largest cohort to date (15). The study by Chalayon 1. Aaron AD. Ankle fusion: a retrospective review. Orthopedics 13:1249–1254, et al (15) reported a 6.4% nonunion rate for their anterior plate cohort 1990. (3 of 47), a similar finding to our 2.1% nonunion rate (1 of 47), 2. Agel J, Coetzee JC, Sangeorzan BJ, Roberts MM, Hansen ST Jr. Functional limitations considering the sample sizes (15). However, despite a large amount of of patients with end-stage ankle arthrosis. Foot Ankle Int 26:537–539, 2005. detailed pertinent information, their study was unclear regarding 3. Coester LM, Saltzman CL, Leupold J, Pontarelli W. Long-term results following ankle arthrodesis for post-traumatic arthritis. J Bone Joint Surg Am 83:219–228, whether the anterior plate fixation constructs consisted of locking 2001. technology, single column versus double column plating, and/or used 4. Friedman RL, Glisson RR, Nunley JA II. A biomechanical comparative analysis of – the lag screw technique in conjunction with the anterior plate two techniques for tibiotalar arthrodesis. Foot Ankle Int 15:301 305, 1994. 5. Haddad SL, Coetzee JC, Estok R, Fahrbach K, Banel D, Nalysnyk L. Intermediate and construct. Their study found that nonunion was 3 times more likely in long-term outcomes of total ankle arthroplasty and ankle arthrodesis: a system- patients with previous subtalar fusion and 2 times more likely in atic review of the literature. J Bone Joint Surg Am 89:1899–1905, 2007. those with preoperative varus malalignment (15). 6. Kitaoka HB. Arthrodesis of the ankle: technique, complications, and salvage treatment. Instr Course Lect 48:255–261, 1999. Our study had some limitations. Variability existed in the crossed 7. Moeckel BH, Patterson BM, Inglis AE, Sculco TP. Ankle arthrodesis: a comparison of screw group, including the various incisional approaches and specific internal and external fixation. Clin Orthop Relat Res 268:78–83, 1991. 4 M.A. Prissel et al. / The Journal of Foot & Ankle Surgery xxx (2017) 1–4

8. Chen YJ, Huang TJ, Shih HN, Hsu KY, Hsu RW. Ankle arthrodesis with cross screw 12. Tochigi Y, Suh JS, Amendola A, Pedersen DR, Saltzman CL. Ankle alignment on fixation: good results in 36/40 cases followed 3-7 years. Acta Orthop Scand lateral radiographs. Part 1: sensitivity of measures to perturbations of ankle 67:473–478, 1996. positioning. Foot Ankle Int 27:82–87, 2006. 9. Mohamedean A, Said HG, El-Sharkawi M, El-Adly W, Said GZ. Technique and short- 13. Tochigi Y, Suh JS, Amendola A, Saltzman CL. Ankle alignment on lateral radio- term results of ankle arthrodesis using anterior plating. Int Orthop 34:833–837, 2010. graphs. Part 2: reliability and validity of measures. Foot Ankle Int 27:88–92, 2006. 10. Monroe MT, Beals TC, Manoli A II. Clinical outcome of arthrodesis of the ankle 14. Veljkovic A, Norton A, Salat P, Saltzman C, Femino J, Phisitkul P, Amendola A. using rigid internal fixation with cancellous screws. Foot Ankle Int 20:227–231, Lateral talar station: a clinically reproducible measure of sagittal talar position. 1999. Foot Ankle Int 34:1669–1676, 2013. 11. Clifford C, Berg S, McCann K, Hutchinson B. A biomechanical comparison of in- 15. Chalayon O, Wang B, Blankenhorn B, Jackson JB III, Beals T, Nickisch F, Saltzman CL. ternal fixation techniques for ankle arthrodesis. J Foot Ankle Surg 54:188–191, Factors affecting the outcomes of uncomplicated primary open ankle arthrodesis. 2015. Foot Ankle Int 36:1170–1179, 2015.