Treatment of malunited fractures of the ankle

Treatment of malunited fractures of the ankle A LONG-TERM FOLLOW-UP OF RECONSTRUCTIVE SURGERY I. I. Reidsma, P. A. Nolte, R. K. Marti, E. L. F. B. Raaymakers From Academic Medical Center, Amsterdam, Netherlands In this retrospective study, using the prospectively collected database of the AO- Documentation Centre, we analysed the outcome of 57 malunited fractures of the ankle treated by reconstructive osteotomy. In all cases the position of the malunited fibula had been corrected, in several cases it was combined with other osteotomies and the fixation of any non-united fragments. Patients were seen on a regular basis, with a minimum follow-up of ten years. The aim of the study was to establish whether reconstruction improves ankle function and prevents the progression of arthritic changes. Good or excellent results were obtained in 85% (41) of patients indicating that reconstructive surgery is effective in most and that the beneficial effects can last for up to 27 years after the procedure. Minor post-traumatic arthritis is not a contraindication but rather an indication for reconstructive surgery. We also found that prolonged time to reconstruction is associated negatively with outcome. I. I. Reidsma, MD, PhD, Resident Department of Radiology Leiden University Medical Center, Albinusdreef 2, 2333 ZA Leiden, The Netherlands. P. A. Nolte, MD, PhD, Orthopaedic Surgeon Department of Orthopaedic Surgery Spaarne Hospital, Spaarnepoort 1, 2134 TM Hoofddorp, The Netherlands. R. K. Marti, MD, PhD, Professor E. L. F. B. Raaymakers, MD, PhD, Orthopaedic Surgeon Department of Orthopaedic Surgery G4-263 Academic Medical Center, University of Amsterdam, Meibergdreef, 115 AZ Amsterdam, The Netherlands. Correspondence should be sent to Dr E. L. F. B. Raaymakers; e-mail: e.l.raaymakers@amc.uva.nl 21 British Editorial Society of Bone and Joint Surgery doi:1.132/31-62x.92b1. 2254 $2. J Bone Joint Surg [Br] 21;92-B:66-7. Received 4 March 29; Accepted after revision 7 September 29 Shortening of the fibula may follow an ankle fracture, particularly when the fracture of the fibula is oblique or spiral. This includes Weber B, but mainly Weber C, fractures. Fractures in the proximal half of the fibula, the Maisonneuve type, are frequently missed and may result in malunion. Shortening of the fibula is a consequence of injury to the ligaments of the syndesmosis, leading to an unstable diastasis. Rupture of the deltoid ligament or a fracture of the medial malleolus is a logical consequence of the mechanism of injury and is responsible for the lateral shift and valgus tilt of the talus, which no longer moves smoothly between the malleoli. This may result in a poor outcome and ultimately lead to post-traumatic arthritis. Early degenerative arthritis will develop, as lateral displacement of the talus by only 1 mm will produce a 42% reduction in the area of tibiotalar contact, thereby increasing the stress on the articular cartilage. 1 Therefore, even a minimal degree of shortening of the fibula requires operative treatment, the aim of which is to reconstruct joint congruency. Restoration of length with correction of the axis and rotational deformity of the fibula is an absolute priority. 2,3 Only then will the lateral malleolus fit anatomically into the fibular notch, which is the key for a successful outcome (Figs 1 and 2). In 199 we reported that reconstructive osteotomies of the fibula in malunited fractures of the ankle had a good clinical outcome at a mean follow-up of five years. 4 Because this cohort consisted of relatively young patients with a mean age of 38 years (2 to 63), we wished to determine the long-term results of reconstructive surgery in these patients. We now present the results of a group with a minimum follow-up of ten years. We have also determined the risk factors associated with a poor clinical outcome. Patients and Methods Between 1975 and 1997, 57 patients with malunited fractures of the ankle who experienced pain and impaired function underwent reconstructive surgery of the fibula with, if necessary, correction of the medial malleolus. Patients who already had major post-traumatic arthritic changes or severe impairment of function were not eligible for operation. The procedure involved lateral capsulectomy and partial resection of the scar tissue at the syndesmosis and, if necessary, medial capsulectomy and debridement. The most important step was a transverse fibular osteotomy and correction of the position of the lateral malleolus in the fibular notch after lengthening, using a small laminar spreader or a compression/ distraction apparatus. 2,3 The lengthened fibula was stabilised with a small plate and the defect filled with cancellous bone graft taken from the distal tibial metaphysis. If the mortise of the 66 THE JOURNAL OF BONE AND JOINT SURGERY

TREATMENT OF MALUNITED FRACTURES OF THE ANKLE 67 Fig. 2a Fig. 2b Fig. 1a Fig. 1b Radiographs showing a) typical deformity after shortening of the fibula: talar shift and tilt and b) reduction of the talus after lengthening of the fibula using the AO compression-distraction apparatus. ankle was unstable, congruency was obtained with a positioning screw, a full threaded screw through the fibula and tibia 5 cm to 7 cm proximal to the ankle joint line which fixes both bones temporarily together. The congruence and function of the ankle were tested during the operation and per-operative radiological evaluation was performed. After operation the ankle was immobilised in a removable splint for five days so that active and passive exercises could be undertaken until the normal range of movement was achieved. The ankle was then immobilised for six weeks in a below-knee walking cast, with full weight-bearing if the reconstruction was stable without a positioning screw. After multiple procedures or the use of a positioning screw, a nonweight-bearing cast was used for six weeks, followed by a walking cast for two weeks. 5 Patients were followed up regularly for a mean of 15.5 years (1 to 27). Assessment. We modified Weber s protocol for rating the results of treatment, 6 using the pre- and post-operative scores for pain, walking and activity as a subjective outcome and for ankle function, subtalar function and the radiologiocal features as an objective outcome (Table I). Considering the radiological outcome, any progression or signs of arthritis, such as osteophyte formation, narrowing of the joint space and subchondral sclerosis, was recorded, comparing the situation before and after operation. Anatomical restoration was defined by the presence of a normal ankle Shenton s line without any talar tilt or lateral shift. 2 A subjective result was considered excellent when there was no pain, normal walking and full activity without limitation at work or in sport. excellence required full ankle and subtalar function, equal to the uninjured side. An excellent radiological result was defined as an anatomical configuration of the ankle mortise and no progression of arthritis. Fig. 2c Fig. 2d Radiographs taken eight months after injury showing there is a nonunion of the medial malleolus, a nonunion of the anterior syndesmotic avulsion, and the typical talar shift and tilt caused by shortening of the fibula. Anterior syndesmotic ligaments are still attached to the lateral malleolus. b) Osteotomy of the fibula, mobilisation and anatomical reduction of the avulsed syndesmotic fragments leads to lengthening of the fibula and to a normal lateral ankle compartment. Lag screw fixation of the reduced pseudarthrotic medial malleolus completes the reconstruction. c) After ten and d) 19 years of follow-up there has been no progression of the preoperative post-traumatic arthritis and normal ankle function. A good subjective result implied slight pain with excessive use, normal function in work-related activities, but restriction of strenuous activities. An objectively good result meant loss of ankle movement of no more than 1, and slight diminution of subtalar movement compared with the opposite side, without the progression of arthritis. A subjectively fair result implied less pain than before surgery, some improvement of walking, and unchanged activity. A fair objective result meant that joint function was the same or only slightly less than before operation, and that there was some progression of arthritis. A subjectively poor result was defined as more pain with less activity and walking than before operation. Considerable limitation of movement and progression of arthritis denoted a poor objective result. Statistical analysis. This was performed using the Mann- Whitney test to determine differences in the time between the initial trauma and reconstruction in the subgroups of patients. A p-value <.5 was considered to represent a significant difference. VOL. 92-B, No. 1, JANUARY 21

68 I. I. REIDSMA, P. A. NOLTE, R. K. MARTI, E. L. F. B. RAAYMAKERS Table I. Modified Weber rating scale Elements of the scale Pain None Slight pain with excess activity 1 Mild pain with normal activity 2 Pain with standing 3 Pain at rest 4 Walking Normal Restricted in strenuous activities 1 Slight limp 2 Partially disabled 3 Totally disabled 4 Activity Full at work and sport Can work normally but is restricted in some activities 1 Normal work but very limited in activity 2 Partially disabled 3 Totally disabled, must change job 4 Radiographs Anatomical restoration and no progression of arthritis Slight progression 1 Considerable progression 2 Ankle joint function Full, equal to other side Loss of movement 1 or less 1 Loss of motion > 1 but dorsiflexion of 95 possible 2 5 dorsiflexion possible 3 Stiff ankle 4 Subtalar joint function Full, equal to other side Slight diminution 1 Limitation < 5% compared with other side 2 Limitation > 5% 3 No movement 4 Results Of 57 patients subjected to reconstructive surgery, nine were excluded from the final analysis. Five had died before reaching follow-up of ten years. One had emigrated and could not be reached for final follow-up. Another had a below-knee amputation for vascular insufficiency and two had required an arthrodesis. Of the nine patients who were excluded, seven were part of the cohort described in 199. 4 In total 48 patients were available for evaluation of the clinical, functional and radiological outcome after a minimum follow-up of ten years after reconstruction of the ankle. Of the 48 patients, 27 were male and 21 female; their mean age at the time of reconstructive surgery was 34.7 years (17 to 64). Using Weber s classification 38 had type C fractures, seven of which were Maisonneuve type, and ten were type B. The primary treatment was conservative in 23 patients, operative in 19, and in six the fracture was overlooked. The interval between the injury and reconstruction was a mean of 21.8 months (1 to 17). The following operations were performed: lengthening only of the fibula in 26, lengthening with correction of the angulation or rotational deformity in 16, and correction of the angular and/or rotational deformity in six. Additional procedures at the medial side included debridement in six, osteotomy in seven, partial or complete removal of the medial malleolus in six, (re)osteosynthesis in five, and exploration and removal of bony fragments in five. In one patient an osteotomy of the posterior malleolus was performed. A total of 16 patients already had slight arthritic changes before the reconstruction. Although there was no evidence of such changes in 3 patients, the typical signs of a malunion were present, with shortening of the fibula and talar tilt or shift, leading to a typical post-traumatic flat foot. The pre-operative radiographs were not available for evaluation in two patients, but at the time of final follow-up one had no arthritic changes and in the other they were minimal. At the final follow-up only the progression of the arthritic changes was determined, comparing the situation before and after reconstruction. There was no evidence of progression in 23 patients, 14 of whom had no arthritic change, 21 showed slight progression, and four who had incongruence of the ankle mortise at the final follow-up showed considerable progression. Of these patients one had no pre-operative evidence of cartilage damage, whereas three had slight pre-operative changes. Three of these patients had a good subjective result after surgery, and one had a fair result. According to the objective results, three of these patients had a fair to poor result and one a good objective result, with joint function after surgery. There was no correlation between the progression of arthritic changes and the duration of follow-up. The clinical results after a minimum follow-up of at least ten years are shown in Table II and are compared with the data published in 199. 4 It is clear that the long-term results are very satisfactory: in the extended cohort, 41 patients (85%) had a subjective good to excellent result after at least ten years, compared with 22 (71%) in the cohort of 199 who were followed up for five years. good to excellent results were seen in 42 patients (88%) in the current analysis, compared with 24 (77%) in 199. Although the majority of patients experienced a prolonged good to excellent response after reconstructive surgery, seven patients (15%) had a fair subjective response and six patients (12%) a fair to poor objective response. The two features that were found to be associated with a poor response were the length of the interval between the trauma and reconstructive surgery, and the existence of arthritic changes prior to operation. Interval between trauma and reconstructive surgery. Considering the objective data (Fig. 3a), there was a THE JOURNAL OF BONE AND JOINT SURGERY

TREATMENT OF MALUNITED FRACTURES OF THE ANKLE 69 Table II. The subjective and objective results of reconstructive surgery are shown for the 31 patients published in 199, with a follow-up period of five years, for the long-term follow-up results of these patients, 1 and for those of an extended cohort of 48 patients with > 1 years follow-up. 2 Numbers of patients per group are shown, as well as the percentage of the total cohort Grade 199 199 1 Extended 2 199 199 1 Extended 2 Excellent (%) 13 (42) 7 (26) 13 (27) 14 (45) 7 (26) 9 (19) Good (%) 9 (29) 14 (52) 28 (58) 1 (32) 16 (59) 33 (69) Fair (%) 4 (13) 6 (22) 7 (15) 6 (19) 4 (15) 5 (1) Poor (%) 5 (16) () () 1 (3) () 1 (2) p =.12 Interval between trauma and reconstructive surgery (mths) 12 4 2 p =.3 Poor/Fair Good Excellent Fig. 3a Pre-operative arthritic changes (%) 4 2 Poor/Fair Fig. 4 Good Excellent Interval between trauma and reconstructive surgery (mths) 12 4 2 p =.48 p =.16 Fair Good Excellent Fig.3b The interval between trauma and reconstructive surgery in months is shown for individual patients in a) the objective and b) the subjective analysis. Patients are grouped according to their response to reconstructive surgery at final follow-up. analysis: Median interval for patients with a poor/fair result: 3 months; with a good result: 12 months; with an excellent result: seven months. analysis: Median interval to reconstructive surgery for patients with a poor response: 53 months; with a good response: 15 months; with an excellent response: eight months. significantly longer intervalbetween trauma and reconstructive surgery for patients with poor/fair results than for those with good (median 3 versus 12 months, p =.3) Bar chart showing pre-operative arthritic changes for: objective analysis: Percentage of patients with pre-operative arthritic changes for the group with a poor/fair result (83%), the group with a good result (3%) and the group with an excellent result (11%) and subjective analysis: Percentage of patients with pre-operative arthritic changes for the group with a poor/ fair result (43%), the group with a good result (39%) and the group with an excellent result (15%). and excellent (median 3 versus seven months, p =.12) results. With the subjective parameters (Fig. 3b), the time to surgery was significantly increased in patients with a fair result compared with those with an excellent result (median 53 versus eight months, p =.16), and those with a fair result compared with those with a good result (median 53 versus 15 months, p =.48). The interval between trauma and reconstructive surgery was significantly longer for the patients described in 199 4 (median 15 months, n = 27) than for those included after 199 (median eight months, n = 21) (p =.4). Pre-operative arthritic changes. In the objective analysis the percentage of arthritic features before reconstructive surgery was highest in the patients with poor/fair results (five of six patients, 83%), less in those with a good response (ten of 33 patients, 3%) and lowest in those with an excellent response (one of nine patients, 11%) (Fig. 4). The subjective results showed less difference in the percentage of pre-operative arthritic changes between patients with a fair response than in VOL. 92-B, No. 1, JANUARY 21

7 I. I. REIDSMA, P. A. NOLTE, R. K. MARTI, E. L. F. B. RAAYMAKERS those with a good response (three of seven patients (43%) vs 11 of 28 patients (39%) than in the objective analysis (Fig. 4). Patients with arthritic changes before operation had a 47% longer interval between trauma and reconstructive surgery than those without pre-operative arthritis (median 17 vs nine months), which appeared to be borderline significant (p =.53). There was no significant difference in age between the subgroups of patients. Discussion Several papers have been published on the outcome of reconstructive surgery after malunion of fractures of the ankle. 2,4,7-9 All emphasise that maximum recovery depends on anatomical restoration of the ankle mortise by adequate lengthening and correction of angular and rotational deformities of the fibula. Most of these papers present good results of reconstruction in most patients, but the cohorts are often small and the periods of follow-up relatively short. This study followed up the patients for at least ten years. We obtained a good or excellent clinical outcome in approximately 85%. These results are better than those reported in our study in 199. 4 However, three of the five patients with a poor response in 199 were excluded from the follow-up as two had an arthrodesis and the other had died. The other two patients had improved. One had been operated on with a modified Kelly technique because of subluxation of the peroneal tendon. 1 Despite these excellent clinical results, arthritic changes had progressed in 52% of the patients at the long-term follow-up. Slight arthritic changes are clearly well tolerated as long as the malalignment of the fibula is corrected and the congruence of the ankle joint is re-established. Poor clinical outcome was significantly associated with a prolonged interval between the accident and reconstructive surgery, as observed before by Offierski et al, 7 indicating that early intervention is essential for malalignment of the ankle mortise after primary management. There was a much higher frequency of pre-operative arthritic changes in the patients with a poor or fair result than in those with a good result. However, arthritic changes do not always influence the clinical outcome as long as perfect alignment of ankle, hind- and midfoot is restored. The existence of minor post-traumatic arthritis is not a contraindication, but rather an indication for reconstructive surgery. We recommend secondary reconstruction in all cases of malunited fractures of the ankle with reasonable function even in the presence of minor arthritic changes. Reconstructive surgery should be performed as early as possible in order to obtain good to excellent long-term clinical results. Other procedures, such as arthrodesis or total ankle replacement, should be considered only if reconstruction fails. No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article. References 1. Ramsey PL, Hamilton W. Changes in tibiotalar area of contact caused by lateral talar shift. J Bone Joint Surg [Am] 1976;58-A:356-57. 2. Weber BG, Simpson LA. Corrective lengthening osteotomy of the fibula. Clin Orthop 1985;199:61-7. 3. Weber BG. Lengthening osteotomy of the fibula to correct a widened mortice of the ankle after fracture. Int Orthop 1981;4:289-93. 4. Marti RK, Raaymakers EL, Nolte PA. Malunited ankle fractures: the late results of reconstruction. J Bone Joint Surg [Br] 199;72-B:79-13. 5. Marti RK, van Heerwaarden RJ, eds. Osteotomies for posttraumatic deformities. Stuttgart: AO Publishing, Thieme Verlag, 28:617-49. 6. Hughes JL, Weber H, Willenegger H, Kuner EH. Evaluation of ankle fractures: non-operative and operative treatment. Clin Orthop 1979;38:111-19. 7. Offierski CN, Graham JD, Hall JH, Harris WR, Schatzker JL. Late revision of fibular malunion in ankle fractures. Clin Orthop 1982;171:145-9. 8. Chu A, Weiner L. Distal fibula malunions. J Am Acad Orthop Surg 29;17:22-3. 9. Sinha A, Sirikonda S, Giotakis N, Walker C. Fibular lengthening for malunited ankle fractures. Foot Ankle Int 28;29:1136-4. 1. Marti RK. Dislocation of the peroneal tendons. Am J Sports Med 1977;5:19-22. THE JOURNAL OF BONE AND JOINT SURGERY