Arthrodesis Versus ORIF for Lisfranc Fractures

n Feature Article

Shahin Sheibani-Rad, MD, MS; J. Christiaan Coetzee, MD; M. Russell Giveans, PhD; Christopher DiGiovanni, MD

abstract Full article available online at Healio.com/Orthopedics. Search: 20120525-26

The Lisfranc joints make up the bony structural support of the transverse arch in the midfoot and account for approximately 0.2% of all fractures. Early recognition and treatment of this injury are paramount to preserving normal foot biomechanics and function. Controversy exists regarding the optimal treatment of patients with Lisfranc injuries, particularly when the instability is entirely ligamentous. A The authors performed a qualitative, systematic review of the literature to compare the 2 most common procedures for Lisfranc fractures: primary arthrodesis and open reduction and internal fixation (ORIF). Six articles with a total of 193 patients met the inclusion criteria. At 1-year follow-up, the mean American Orthopaedic Foot and Ankle Society score of ORIF patients was 72.5 and of arthrodesis patients was 88.0. Fisher’s exact test revealed no significant effect of treatment group on the percentage on patients who had an anatomic reduction (P5.319). B

This study highlights that both procedures yield satisfactory and equivalent results. A slight advantage may exist in performing a primary arthrodesis for Lisfranc joint injuries in terms of clinical outcomes.

C Figure: Myerson classification. Total incongruity (A). Partial incongruity (B). Divergent (C). Dr Sheibani-Rad is from the Department of Orthopaedic Surgery, McLaren Regional Medical Center, Michigan State University, Flint, Michigan; Drs Coetzee and Giveans are from Minnesota Sports Medi- cine and Twin Cities Orthopedics, Minneapolis, Minnesota; and Dr DiGiovanni is from the Department of Orthopaedic Surgery, Brown Medical School, Rhode Island Hospital, Providence, Rhode Island. Drs Sheibani-Rad, Coetzee, Giveans, and DiGiovanni have no relevant financial relationships to disclose. Correspondence should be addressed to: Shahin Sheibani-Rad, MD, MS, Department of Orthopae- dic Surgery, McLaren Regional Medical Center, Michigan State University, 401 S Ballenger Hwy, Flint, MI 48532 ([email protected]). doi: 10.3928/01477447-20120525-26

e868 Healio.com The new online home of ORTHOPEDICS | Healio.com/Orthopedics lisfranc Fractures | Sheibani-Rad et al

he Lisfranc joints make up the bony structural support of the transverse Tarch in the midfoot. Lisfranc injuries account for approximately 0.2% of all fractures,1 with an incidence of 1/55,000 cases per year.1,2 Early recognition and treatment of this injury are paramount to preserving normal foot biomechanics and function. Approximately 20% of Lisfranc joint injuries are missed on initial antero- posterior and oblique radiographs,3 and 1 persistent malalignment after this injury Figure 1: Graph showing flow of studies retrieved for systematic review. can lead to chronic pain, progressive de- formity, and prolonged disability.4,5 Controversy exists regarding the op- tarsocuneiform joint, talometatarsal an- Quenu and Kuss15 or a modification of timal treatment of patients with Lisfranc gle, and the line tangential to the medial Myerson et al5 or Hardcastle et al,16 (2) injuries, particularly when the instabil- aspect of the navicular and the medial cu- treatment within 48 hours of injury, (3) ity is entirely ligamentous.6 Although neiform intersecting the base of the first the use of 1 of the 2 treatments of interest, the current trend is treating these patients metatarsal. Clinical outcome assessment and (4) publication between January 1985 with open reduction and internal fixation was determined via patient scoring based and September 2011 in English. All injury (ORIF),6-8 those with dislocation in the on the postoperative American Ortho- mechanisms were eligible for the study. absence of significant fracture continue paedic Foot and Ankle Society (AOFAS) Exclusion criteria included studies that to pose a therapeutic challenge. Despite midfoot scale. The purpose of the study focused on a biomechanical model, review appropriate initial treatment, up to 94% was to compare clinical and radiographic articles, articles on the pediatric popula- of these patients develop posttraumatic outcomes between arthrodesis and ORIF tion, isolated case reports, surgical tech- arthritis and require a conversion to an ar- interventions for Lisfranc injuries based nique articles, foreign language reports, throdesis of the tarsometatarsal joints.9-11 on available clinical and radiologic out- studies using either closed reduction or Some studies report primary arthrodesis comes. percutaneous pinning as the primary treat- to be a salvage method,12,13 whereas oth- ment, reports of arthrodesis as a salvage ers recommend it as a primary treatment.14 Literature Search or secondary procedure, and patient as- Because no consensus exists for manag- A comprehensive search was per- sessment that did not include the AOFAS ing Lisfranc fracture-dislocation variants, formed of Medline, Cochrane Central midfoot score. All study designs were the authors of the current study performed Register of Controlled Trials, and Ovid eligible, including randomized, controlled a qualitative, systematic review of the lit- using the key words Lisfranc, Lisfranc trials and prospectively or retrospectively erature to compare the 2 most common fracture, Lisfranc dislocation, metatarsal nonrandomized, controlled trials. procedures for Lisfranc fractures: primary fracture, metatarsal dislocation, Lisfranc arthrodesis and ORIF. Review of radio- arthrodesis, Lisfranc ORIF, and Lisfranc Statistical Methods graphic reduction and patient outcomes open reduction internal fixation. The data- Statistical analysis was performed was peformed. base review also included a search of any using SPSS version 16.0 for Windows published report in the American Associa- software (SPSS, Inc, Chicago, Illinois). Materials and Methods tion of Orthopaedic Surgeons (AAOS) an- Fisher’s exact test was performed to com- A formal review was performed of all nual meeting proceedings between 2006 pare the reduction rate between the ORIF available reports in the English literature and 2010. To ensure that all possible ar- group and the arthrodesis group. Statisti- pertaining to ORIF and primary arthrod- ticles were considered, the references of cal significance was set atP ,.05. esis (complete and partial) of the Lisfranc all relevant articles were manually cross- joints. Radiographic outcome assessment referenced to identify potential additional Results included final or postoperative anatomic studies. The database search resulted in 1743 ar- reduction based on >1 of the following For a study to satisfy the inclusion ticles (Figure 1). After limiting the search radiological findings: diastasis between criteria, it required (1) inclusion of a Lis- to clinical trials in English and excluding the first and second metatarsals or meta- franc classification system according to biomechanical and surgical techniques,

JUNE 2012 | Volume 35 • Number 6 e869 n Feature Article

This left a total of 6 studies for the Table 1 current review (Table 1). The studies in- Selected Articles From Systematic Review cluded 8 treatment arms and a total of 193 eligible patients. Of the 6 studies, 4 Author Study Design Fixation Type No. of Patients True Na focused on ORIF alone. Two studies di- Arntz et al6 PCS ORIF 40 40 rectly compared ORIF and arthrodesis. Mulier et al12 PCS ORIF vs primary 28 28 The ORIF group comprised a total 160 arthrodesis eligible patients, with 152 patients under- Kuo et al17 RR ORIF 48 42 going the procedure and returning for fol- Ly & Coetzee18 PRC ORIF vs primary 41 41 low-up. The arthrodesis group comprised arthrodesis 33 patients. Rajapakse et al19 RR ORIF 25 17 Six articles included data on post- Teng et al20 RR ORIF 11 11 operative anatomic reduction,6,12,17-20

Abbreviations: ORIF, open reduction and internal fixation; PCS, prospective consecutive and 4 articles provided AOFAS midfoot series; PRS, prospective randomized comparison; RR, retrospective review. scores.17-20 Of 3 prospective studies, 2 in- aOnly patients who were available for follow-up and/or met inclusion/exclusion criteria, not total patient population. cluded direct comparisons of ORIF and arthrodesis. One study used a randomized design, and in the remaining prospective Table 2 studies, it was unclear whether the patients were randomized. Mean follow-up Comparison of AOFAS Scores Between Selected Studies was 39.8 months.

Kuo Ly & Rajapakse Teng Mean Of the 6 articles that met the inclusion Procedure et al17 Coetzee18 et al19 et al20 AOFAS Significance criteria, 6,12,17-20 4 reported AOFAS scores 17-20 ORIF 80.2 68.6 70.3 71.0 72.5 NA at 1-year follow-up. Each of these arti- Arthrodesis NA 88.0 NA NA 88.0 NA cles included patients treated with ORIF, with 1 of the 4 also reporting AOFAS Abbreviations: AOFAS, American Orthopaedic Foot and Ankle Society; NA, not applicable; ORIF, open reduction and internal fixation. scores for an arthrodesis group (Table 2).18 The total number of patients evaluated was 90 in the ORIF group and 21 in the Table 3 arthrodesis group. At 1-year follow-up, Comparison of Anatomic Reduction the mean AOFAS score of ORIF patients was 72.5 and of arthrodesis patients was Patients ORIF Arthrodesis P 88.0. By not having access to the origi- Total No. 146 33 nal data sets, the authors were unable to perform a statistical analysis on AOFAS No. with reduction 134 28 scores at 1 year between patients under- % 91.8 84.8 .319 going ORIF and arthrodesis. However, al- Abbreviation: ORIF, open reduction and internal fixation. though the sample size is somewhat small for the arthrodesis group, a clinical difference between the 2 groups is evident at case reports, imaging studies, studies not this could not be determined from the ab- 1-year follow-up, with the AOFAS score involving ORIF or arthrodesis, studies not stract. From this group, 16 articles were for the arthrodesis group higher than that involving the Lisfranc joint, nonoperative determined to fulfill the inclusion criteria. of the ORIF group. treatment, and duplicate reports, a total Four articles were rejected because they In addition, the rate of postoperative of 154 applicable reports were identified. combined .1 method of treatment. Three radiographic anatomic reduction was also The abstracts from the 154 articles were studies were rejected because they did not evaluated from the data reported by the 6 examined by an author (S.S.R.) to deter- specify their requirement for anatomical re- articles that met the inclusion criteria (Ta- mine whether the article fit the inclusion duction on radiographs. Three studies were ble 3). Combining the patient totals from or exclusion criteria. Each original article rejected because they used patient assess- these articles, a total of 146 patients were was reviewed to determine eligibility if ment other than the AOFAS midfoot score. treated with ORIF. Of these, 134 (91.8%)

e870 Healio.com The new online home of ORTHOPEDICS | Healio.com/Orthopedics lisfranc Fractures | Sheibani-Rad et al

patients had postoperative anatomic re- Postoperative Care duction. Two of the 6 articles reported All studies preferred partial or non- Table 4 data on patients treated with arthrodesis. weight bearing for at least 6 to 8 weeks Mechanism of Injury by A total of 33 patients were included, and postoperatively. The K-wires in the lat- Percentage of Occurrence 28 (84.8%) had postoperative anatomic eral column were usually removed 6 to reduction. Fisher’s exact test revealed no 8 weeks postoperatively. Several authors Mechanism of Injury Occurrence, % significant effect of treatment group on advocated maintenance of cortical screws Motor vehicle accident 57 the percentage on patients who had an in the medial column unless they were a Fall from height 21 6,20 anatomic reduction (P5.319). source of pain, whereas some preferred Sports related 9 Patient demographics were not de- to remove the screws between 18 weeks Work related 8 scribed in all studies. Of the 193 patients and 6 months postoperatively.12,17-19 Crush injury 5 across all studies, 127 were men and 66 were women. In studies in which the age Complications of the patient population was reported, Across all studies, the most common Table 5 mean overall age was 29 years (range, complications reported were screw break- 15-76 years). The most common cause of age, posttraumatic arthritis, missed asso- Complications injury was high-energy trauma, usually ciated injuries, deep venous thrombosis, a motor vehicle accident (Table 4). The compartment syndrome, and superficial ORIF Group Arthrodesis Group most common injury type was B2 accord- wound infection (Table 5). Screw breakage Pseudarthrosis ing to Myerson et al’s5 classification (Fig- Posttraumatic OA Painful hardware ure 2). More than half of the injuries were Discussion DVT Nonunion mixed osseous and ligamentous, and the Prompt recognition and treatment of Compartment Delayed union remaining injuries were ligamentous. tarsometatarsal injury is imperative to syndrome minimize the potential for significant long- Superficial infection Surgical Technique term disability. A high index of suspicion Nearly all studies included in the cur- is warranted for these injuries because they Missed associated injury rent review used a dorsal incision tech- are frequently subtle or occult and can Abbreviations: DVT, deep venous nique over the first intermetatarsal space to therefore be easily missed. When suspicion thrombosis; OA, osteoarthritis; ORIF, expose the medial column joint surfaces of is present despite the absence of identifi- open reduction and internal fixation. the midfoot. Comminuted fractures were able abnormality on plain radiograph as- then reduced, and smaller fragments were sessment, weight-bearing radiographs and removed. The medial border of the base of stress view examinations are recommend- of interposed soft tissue structures that of- the first metatarsal and the medial border ed. General agreement exists in the litera- ten impede anatomic reduction.16,24 Long- of the first cuneiform were aligned. The ture that anatomic reduction of the Lisfranc term outcome after Lisfranc injury appears base of the second metatarsal was opposed joint is important for optimal outcome.4,9,21 to be related to the accuracy of reduction. to the lateral border of the first cuneiform. Adib et al22 found a significant difference The current review is the first systematic The joint was held reduced with provi- between anatomic reduction and the devel- review of literature comparing both ana- sional Kirschner wires, then one 3.5- or opment of osteoarthritis. In patients with tomic reduction and clinical outcomes with 4.5-mm transarticular countersunk corti- anatomic reduction, 35% developed osteo- respect to ORIF and primary arthrodesis in cal screw was inserted from the metatarsal arthritis, compared with 80% of those who Lisfranc injuries. A previous systematic re- base proximally into the medial cuneiform. had nonanatomic reduction.22 view of the literature has been performed, If the third metatarsal base was dislocated, However, no general consensus exists with several limitations.25 Unlike the pre- a second dorsal incision was typically as to the best method of treatment for pa- vious review, the current review stratified performed between the third and fourth tients with Lisfranc injuries, aside from fractures according to a well-accepted clas- metatarsals to stabilize the third metatarso- agreement that closed reduction is unsuc- sification system and compared anatomic cuneiform joint and enable the fourth and cessful in the majority of cases given the reduction and clinical outcomes based on fifth metatarsal bases to be fully reduced. inherent instability associated with this 1 accepted parameter of interest. Depending on the type of fracture, the lat- injury pattern.23 Closed reduction and Open reduction and internal fixation eral (fourth and fifth) tarsometatarsal joints casting provides neither proper restraint is currently the most common treatment were generally stabilized using K-wires. against further displacement nor removal for displaced Lisfranc joint injuries. The

JUNE 2012 | Volume 35 • Number 6 e871 n Feature Article

2A 2B 2C Figure 2: Myerson classification. Total incongruity (A). Partial incongruity (B). Divergent (C). treatment of complete ligamentous inju- The most common complications several authors (S.S.-R., C.D., M.R.G.). ries poses a therapeutic challenge for most across all studies were screw breakage and Despite the strengths of the current study, surgeons. Even with anatomic reduction, posttraumatic arthritis (Table 5). A total weaknesses existed in the literature. First, treatment of these injuries does not pro- of 3 patients were noted to develop com- the AOFAS midfoot score is not a vali- duce consistently satisfactory outcomes. partment syndrome requiring fasciotomy, dated tool. This scale was chosen because Between 40% and 94% of these patients suggesting that this is a somewhat unusual it was the most common method of out- develop posttraumatic arthritis, requiring clinical event after Lisfranc injury. It is come assessment in the literature. Second, conversion to arthrodesis.9,13,17 Previous postulated that the incisions and anatomic some studies did not provide key data el- studies have reported the reoperation rate reduction decompress the midfoot and ements, including individual patient AO- in the ORIF group to be between 75% and decrease the likelihood of compartment FAS scores. An attempt was made to con- 79%, compared with 17% to 20% in the syndrome. One patient in the arthrodesis tact the authors of those studies to retrieve arthrodesis group.10,18 group was found to develop nonunion individual patient scores without success. The variation in the rate of arthrosis requiring revision arthrodesis with bone One study used the AOFAS midfoot score across multiple studies suggests that the graft from the calcaneus; the fusion healed to measure clinical results for patients injury itself, rather than any particular uneventfully.12 It was postulated that the treated with primary arthrodesis.18 Thus, chosen treatment, has the greatest influ- high fusion rate after primary fusion may the current authors were unable to statisti- ence on outcome. In a recent prospective, be due to the hyperemia that follows the cally compare the ORIF and primary ar- randomized study by Henning et al,26 pa- severe injury, although it is suspected that throdesis groups. tients treated with ORIF had a higher rate surgeon experience also plays a role. Simi- Few articles report prospective studies; of additional surgery compared with those larly, a patient with delayed union received 2 articles report the primary arthrodesis treated with primary arthrodesis (79% vs treatment with a bone stimulator and pro- procedure. Given the nature of the current 17%, respectively; P,.05). The primary ceeded to union within 8 weeks.12 Given systematic review, a priori power analysis arthrodesis group trended toward better the fact that the nonunion rate was low in is unjustified. Fisher’s exact test was used Short Musculoskeletal Functional Analy- the arthrodesis group, this may be a fac- to give an exact P value for the measure of sis scores than the ORIF group.26 In the tor in deciding to proceed with primary statistical difference. The relatively small current review, no apparent differences arthrodesis rather than ORIF. Approxi- number of arthrodesis patients (n533) existed between ORIF and primary ar- mately 25% of patients in the arthrodesis still surpasses the acceptable sample size throdesis in terms of anatomic reduction. group developed asymptomatic degenera- of 30 for sufficient power for Fisher’s ex- This review also revealed that the mean tive joint disease, which required no fur- act test, as well as for accepted interpreta- AOFAS score was higher in the primary ther surgery. tion of statistical results. The large sample arthrodesis group at 1-year follow-up for A strength of the current systematic of ORIF patients lends itself to be viewed pure ligamentous and combined bony and review was the clear objective set forth at as a truer representation of the actual ligamentous injuries. The current study’s the start of the study to eliminate selec- population, yet the current authors claim data support primary arthrodesis as a pri- tion bias. The article search using elec- that while comparing this with a group mary treatment for Lisfranc joint injuries tronic databases and manual bibliography of 33 patients, the nonsignificant finding due to a significantly decreased rate of ad- searches was comprehensive and repeated ascertained by Fisher’s exact test gives a ditional surgeries, as well as a tendency by 3 authors (S.S.-R., C.D., M.R.G.). The statistically valid interpretation of the dif- toward improved clinical outcome scores protocol was strictly adhered to, and all fering (or nondiffering, in this case) rates when compared with ORIF. potential articles had to be agreed on by of anatomic reduction.

e872 Healio.com The new online home of ORTHOPEDICS | Healio.com/Orthopedics lisfranc Fractures | Sheibani-Rad et al

Conclusion 7. Curtis MJ, Myerson MS, Szura B. Tarso- 17. Kuo RS, Tejwani NC, DiGiovanni CW, et al. metatarsal joint injuries in the athlete. Am J Outcome after open reduction internal fixa- This study provides evidence that Sports Med. 1993; 21:497-502. tion of Lisfranc joint injuries. J Bone Joint Surg Am ORIF and arthrodesis yield satisfactory 8. Myerson M. The diagnosis and treatment of . 2000; 82(11):1609-1618. and equivalent results. A slight advantage injuries to the Lisfranc joint complex. Orthop 18. Ly TV, Coetzee JC. Treatment of primarily may exist to performing a primary ar- Clin North Am. 1989; 20:655-664. ligamentous Lisfranc joint injuries: primary 9. Buzzard BM, Briggs PJ. Surgical manage- arthrodesis compared with open reduction throdesis for Lisfranc joint injuries in J Bone Joint Surg Am ment of acute tarsometatarsal fracture dis- and internal fixation. . terms of clinical outcomes. Although not location in the adult. Clin Orthop Relat Res. 2006; 88(3):514-520. all patients are candidates for primary ar- 1998; 353:125-133. 19. Rajapakse B, Edwards A, Hong T. A single 10. Henning JA, Jones CB, McDermot M, An- surgeon’s experience of treatment of Lisfranc throdesis, it should be considered for Injury derson JG, Bohay DR. ORIF vs. primary joint injuries. . 2006; 37:914-921. those with complete ligamentous injuries. arthrodesis of lisfranc injuries: a prospective 20. Teng AL, Pinzur MS, Lomasney L, Mahoney Further prospective studies with direct randomized trial. Paper presented at: AAOS L, Havey R. Functional outcome follow- comparisons of the 2 procedures are need- Annual Meeting; February 14-18, 2007; San ing anatomic restoration of tarsal-metatarsal Diego, CA. fracture dislocation. Foot Ankle Int. 2002; ed. In the future, a validated scoring sys- 11. Richter M, Zech S, Geerling J, Frink M, 23(10):922-926. tem such as the Visual Analog Scale Foot Knobloch K, Krettek C. A new foot and 21. Pérez Blanco R, Rodríguez Merchán C, and Ankle11 should be used. ankle outcome score: questionnaire based, Canosa Sevillano R, Munuera Martínez L. subjective,Visual-Analogue-Scale, validated Tarsometatarsal fractures and dislocations. J and computerized. Foot Ankle Surg. 2006; Orthop Trauma. 1988; 2(3):188-194. References 12:191-199. 22. Adib F, Medadi F, Guidi E, et al. Osteoarthri- 1. Aitken AP, Poulson D. Dislocation of the 12. Mulier T, Reynders P, Dereymaeker G, Broos tis following open reduction and internal fix- tarsometatarsal joint. J Bone Joint Surg Am. P. Severe Lisfranc injuries: primary arthrod- ation of the Lisfranc injury. Paper presented 1963; 45:246-260. esis or open reduction and internal fixation? at: 12th EFORT Congress; June 1-4, 2011; 2. English TA. Dislocations of the metatarsal Foot Ankle Int. 2002; 23(10):902-905. Copenhagen, Denmark. bone and adjacent toe. J Bone Joint Surg Br. 13. Sangeorzan BJ, Veith RG, Hansen ST Jr. Sal- 23. Faciszewski T, Burks R, Manaster B. Subtle 1964; 46:700-704. vage of Lisfranc’s tarsometatarsal joint by injuries of the Lisfranc joint. J Bone Joint 3. Rosenburg GA, Patterson BM. Tarsometa- arthrodesis. Foot Ankle. 1990; 10:193-200. Surg Am. 1990; 72:1519-1522. tarsal (Lisfranc’s) fracture dislocation. Am J 14. Coetzee JC, Ly TV. Treatment of primarily 24. Jeffreys T. Lisfranc’s fracture-dislocation: A Orthop (Belle Mead NJ). 1995; (suppl):7-16. ligamentous Lisfranc joint injuries: primary ar- clinical and experimental study of tarso-meta- 4. Goosens M, De Stoop N. Lisfran fracture dis- throdesis compared with open reduction and in- tarsal dislocations and fracture-dislocations. J locations: etiology, radiology, and result of ternal fixation. Surgical technique. J Bone Joint Bone Joint Surg Br. 1963; 45:546-551. Surg Am treatment. A review of 20 cases. Clin Orthop . 2007; 89(suppl 2 pt 1):122-127. 25. Stavlas P, Roberts CS, Xypnitos FN, Gian- Relat Res. 1983; 176:154-162. 15. Meyer SA, Callaghan JJ, Albright JP, Crow- noudis PV. The role of reduction and internal 5. Myerson MS, Fisher RT, Burgess AR, Ken- ley ET, Powell JW. Midfoot sprains in col- fixation of Lisfranc fracture-dislocations: a zora JE. Fracture dislocations of the tarso- legiate football players. Am J Sports Med. systematic review of the literature. Int Or- metatarsal joints: end results correlated with 1994; 22:392-401. thop. 2010; 34(8):1083-1091. pathology and treatment. Foot Ankle. 1986; 16. Hardcastle P, Reschauer R, Kutscha-Lissber 26. Henning JA, Jones CB, Sietsema DL, Bohay 6:225-242. E, Schoffman W. Injuries to the tarsometatar- DR, Anderson JG. Open reduction inter- 6. Arntz CT, Veith RG, Hansen ST. Fractures and sal joint: Incidence, classification and treat- nal fixation versus primary arthrodesis for fracture dislocation of the tarsmetatarsal joint. ment. J Bone Joint Surg Br. 1982; 64:349-356. Lisfranc injuries: a prospective randomized J Bone Joint Surg Am. 1988; 70:173-181. study. Foot Ankle Int. 2009; 30(10):913-922.

JUNE 2012 | Volume 35 • Number 6 e873