Introduction
Ankle fractures were first described by Sir Percival Pott in 1768 and are one of the most common skeletal injuries seen in clinical practice [
1]. It has been estimated that they comprise 9 % of all fractures and up to 22.6 % of all lower-limb fractures in the UK population [
2,
3]. Their prevalence is rising as a consequence of osteoporosis in an increasing elderly population [
4]. An epidemiological study of 1,500 ankle fractures revealed that isolated distal fibular or lateral malleolus fractures occurred in two thirds of patients, whilst bimalleolar fractures occurred in a quarter and trimalleolar fractures in the remaining 7 % [
5].
The two most universally accepted classification schemes are the Danis-Weber and Lauge-Hansen systems [
6‐
8]. Whilst both allow clinicians to define and communicate the fracture pattern, managing these injuries is primarily based on an assessment of stability, which incorporates the amount of displacement, presence of medial injury and associated talar shift [
9,
10]. The treatment aims are to obtain reduction of displaced fractures, maintain anatomic alignment of the ankle mortise and achieve bony union. The closed treatment of stable fractures routinely involves a short period of casting and is usually successfully followed by a progressive return to weight bearing and physiotherapy [
11,
12]. Unstable fractures with disruption of the mortise require open reduction and internal fixation (ORIF). Initially, this involves closed reduction and temporary stabilisation by casting or external fixation. ORIF can be performed safely once soft tissue swelling has settled in order to reduce the risk of wound complications [
13]. Most commonly, ORIF is performed using Arbeitsgemeinschaft für Osteosynthesefragen (AO) principles through the use of compression screws and a neutralisation plate [
14]. Precontoured locking plates have been introduced to allow improved fixation in osteoporotic bone [
15]. However, lateral plating can lead to complications such as wound breakdown and infection due to the poor skin envelope that surrounds the distal fibula [
16]. Furthermore, patients often complain of prominent hardware that require subsequent removal [
17,
18]. A recent systematic review of 1,822 ankle fractures treated with ORIF revealed that approximately one fifth of optimally reduced fractures had unsatisfactory results with regards to functional outcome, subjective outcome and radiographic evaluation [
19].
Intramedullary fixation includes the use of both compression screws and intramedullary nailing (IMN). As mini-incision techniques are used, these techniques may benefit patients with compromised skin by reducing the risk of wound complications. In addition, soft-tissue swelling may not present a contraindication to early fixation, potentially allowing earlier surgery and thus earlier discharge from hospital. However, a formal assessment of the overall outcomes of patients treated with these methods has never been made. Therefore, the aim of this systematic review was to evaluate the clinical and functional results of patients with distal fibular fractures treated with intramedullary devices.
Discussion
Intramedullary fixation is a well-established technique for managing long-bone fractures. Standard AO plating of distal fibular fractures achieves acceptable and consistent union rates but has been associated with wound infection, wound breakdown and hardware prominence, with reported complication rates of up to 30 % [
41‐
43]. Due to the mini-incision technique and low-profile implants associated with intramedullary fixation of distal fibular fractures, there is a theoretical reduction in the risk of patients developing wound complications and soft-tissue irritation due to hardware prominence. The purpose of this systematic review was to evaluate the results of intramedullary fixation with regards union, functional outcome and complications.
Earlier studies evaluated the outcome of distal fibular fractures treated with intramedullary screws [
24,
25,
31]. Bankston et al. [
24] used open reduction techniques and inserted 4.2-mm fully threaded screws in compression mode. Cerclage wires were used at the surgeons’ discretion for improved stability at the fracture site. Ray et al. [
25] specified that the fracture pattern must be transverse and short oblique or minimally comminuted, otherwise it is not possible to maintain fibular length with intramedullary screws. They used closed reduction techniques under image guidance with occasional percutaneous use of a towel clip. An advantage of this method is that patients can be treated on an outpatient basis. Lee et al. [
31] used newer cannulated, headless, variable-pitch screws following open reduction. For comminuted fractures, they recommended the use of cerclage wires or sutures. They reported that the compressive force exerted by headless variable-pitch screws allows enough stability to resist proximal migration and rotation at the fracture site. This may explain why their series gave the maximum union rate, greatest functional outcome and lowest complication rate of the reviewed studies.
Unlocked IMN of distal fibular fractures was first reported by McLennan and Ungersma following the development of their Inyo nail made from malleable stainless steel and triflanged to resist torsional stress [
21]. However, their initial series gave an unacceptably high complication rate of 16 % due to nail migration and malunion. Following refinement of their technique to include percutaneous clamping and the use of shorter nails, they were able to achieve a reduced complication rate of 10.6 %, with no cases of malunion or nail migration [
22]. Pritchett et al. [
23] compared rush rods to traditional AO plating methods but only included supination eversion type IV injuries. They experienced an earlier time to weight bearing in the rush rod group (6 vs. 12 weeks) and more complications (deep infection, nonunion, ankle fusion) in the AO plate group. Whilst improved functional outcomes were reported with rush rods, worse radiographic results were seen in terms of fibular shortening, increased medial clear space and posterior displacement. Lee et al. [
34] compared the use of Knowles pins to plating and noted that the pin group had significantly smaller wound incisions, a shorter operative time, a shorter hospital stay, less symptomatic hardware and lower complication rates. However, this study was limited by its retrospective nature and nonrandomised group allocation. Importantly, there was no significant difference in functional outcomes at final follow-up.
Advantages of locked IMN include better rotational control, improved stability and reduced risk of nail migration. The first report of locked IMN for distal fibular fractures was published by Kara et al. [
26] concerning the ANK nail, which was designed for lateral malleolus fractures with syndesmosis rupture. Whilst all fractures healed, the most significant complication was fibular shortening, which occurred in comminuted, oblique or nonanatomically reduced fractures. Kabukcuoglu et al. [
27] reported limited success with the ANK nail, with an overall complication rate of 20.4 %. They correlated a significantly worse clinical and functional outcome with fibular shortening >2 mm. Ramasamy and Sherry [
28] provided the first report of a modern fibular nail involving patients with Weber B fractures. However, due to a very small sample size, their results have limited external validity. Rajeev et al. [
32] reported on a larger cohort of elderly patients treated with the same implant and type noted that all fractures healed uneventfully with no complications. Functional assessment revealed a mean OMS at 1 year of 58.125, i.e. fair.
Gehr et al. [
29] presented the largest study to date regarding IMN of distal fibular fractures. This prospective case series reduces the possibility of recall bias associated with earlier retrospective studies. In addition, a consecutive group of patients was followed, which helps eliminate the risk of recruitment bias within their study population. Whilst acceptable results were reported, two patients suffered severe soft-tissue complications requiring skin grafting procedures. This may have been due to the use of open reduction techniques in some patients with associated complex nonmalleolar distal tibial fractures. Bugler et al. [
33] reported radiological and functional outcomes of locked IMN in a large series of patients with 6 years follow-up. An independent and blinded assessor was used to interpret radiographic outcomes, potentially eliminating the risk of investigator bias. However, whilst 76.2 % of patients were available for radiographic follow-up, only 49.5 % responded to the postal questionnaires regarding functional outcome, indicating a significant proportion of their population lost to follow-up. All fractures eventually united, and acceptable functional results were achieved through a variety of validated scoring systems [
40,
44,
45]. However, the overall complication rate was 22.8 %. Specifically, complications relating to fixation failure were higher earlier in the series when unstable locking screw configurations were used. Eventually, the combination of syndesmosis and distal locking screws was deemed to be the most stable configuration. Tawari et al. [
36] reviewed two matched groups of patients who underwent fixation for unstable Weber B fractures with either IMN or standard AO plating. There was no significant difference between groups in time taken to achieve clinical and radiological union. One patient in the plate group had a wound infection, but there were no wound complications in the nailing group. Most recently, Bugler et al. [
37] presented the only prospective randomised controlled trial comparing locked IMN to plating. Whilst 16 % of patients in the plating group developed wound infections, no infections or wound complications occurred in the IMN group. At 1 year, functional outcome favoured the IMN group, but this difference was statistically insignificant. In addition, the overall cost of treatment in the IMN group was lower despite the increased cost of the implant.
Overall, union rates were well reported throughout, with bony consolidation being achieved in 98.9 % (range 88.9–100 %) of patients [
21‐
26,
28‐
36]. It can be concluded that intramedullary fixation of distal fibular fractures gives excellent union rates comparable with ORIF. However, methods of assessing union were not accurately presented, making this an irreproducible outcome measure. Typically, plain radiographs are used to assess union of long-bone fractures through the presence of callus formation, but this has shown to be unreliable, with wide interobserver variability [
46]. In practice, clinical evaluation is essential and incorporates an assessment of pain, tenderness and ability to bear weight; details of how these factors were measured were not presented in the reviewed studies. Functional assessment was undertaken through a variety of assessment tools. Earlier studies used unvalidated patient-reported outcome scores, with good or excellent outcomes reported by the vast majority of patients [
21,
22,
25,
26,
30,
31,
35]. The OMS has been used primarily in studies evaluating locked IMN in which 68.8 % achieved good or excellent outcomes [
28,
29,
32,
33]. Whilst these results are encouraging, those with unsatisfactory functional outcomes may reflect the natural history of such injuries, which often occur in elderly patients. The OMS was initially conceived to provide a functional assessment tool following an ankle fracture and was tested against subjective evaluation, range of motion, presence of osteoarthritis and severity of initial injury [
40]. Whilst widely used, it is important to note that limitations of the score’s validity testing include a relatively small series of patients, inclusion of only bimalleolar fractures and scoring questions relating to running and jumping, which many elderly patients would be unable to do prior to injury. In addition, the presence of syndesmotic injuries with the potential for distal tibiofibular joint instability deserves special mention. None of the intramedullary screw devices [
24,
25,
31] and earlier designs of nailing implants [
21‐
23,
26‐
28,
30,
32,
34,
35] allowed for combined fixation of associated syndesmotic injuries. Therefore, these older devices would not be indicated in more complex injuries, such as high fibular fractures. However, more modern locked nailing implants [
29,
33,
36,
37] allow for supplementary syndesmosis screw fixation and are therefore more suited to such injuries.
The mean complication rate across all studies was 10.3 % (range 0–33.3 %) [
21‐
34,
36,
37]. The wide range probably reflects multiple variables, such as different nail design, surgeon experience and complexity of cases. Most commonly, complications involved implant-related problems requiring metalwork removal, failure or fibular shortening [
21‐
37]. The latter being mainly reported with the use of the ANK nail in unsuitable fracture patterns, whilst problems relating to metalwork were associated with either earlier nail design or improper locking screw placement. Interestingly, complications such as mechanical failure were relatively higher in the locked IMN series compared with other studies. Whilst this may appear counterintuitive, it may reflect the learning curve seen in longer studies in which earlier techniques were deemed to be inadequate [
33]. The large variation between studies makes it difficult to accurately compare complication rates to standard plating techniques. The most recent studies involving AO techniques show extremely favourable complication rates of 1.7–5 % [
47,
48]. However, the only prospective randomised controlled study available for review showed significantly more wound-related complications in patients treated with plates than those treated with IMN [
37].
Overall, review of the selected studies revealed that excellent union rates and satisfactory functional outcomes can be expected with intramedullary fixation. However, complication rates can be unacceptably high, although this may reflect a learning curve. Due to the numerous methodological flaws within the reviewed studies, definitive conclusions regarding the clinical application of intramedullary fixation for distal fibular fractures cannot be made. The presence of selection bias, retrospective data collection, lack of control groups and inadequate functional assessment tools provide only poor-quality evidence for fibular nailing. In conjunction with the paucity of high-quality evidence regarding clinical and functional outcomes, practical concerns exist regarding the steep learning curve, expense and appropriate timing of surgery with regards to soft-tissue swelling.
Strengths of this review are the clarity and reproducibility of our search strategy using multiple evidence-based databases. PRISMA guidelines for the reporting of systematic reviews were used throughout in order to increase transparency and reduce the risk of publication bias [
20]. Limitations of this review are the inability to pool data for true meta-analysis due to the heterogeneity of individual studies. Also, most reviewed articles were case series, which are prone to both selection and experimental bias. We also acknowledge that although insufficient detail was available in abstracts to allow complete critical appraisal, they were included in our study in order to provide the most up-to-date assessment regarding locked IMN [
36,
37].
In conclusion, there is insufficient evidence for changing practice from plating of unstable distal fibular fractures to intramedullary fixation based on the current literature. Adequately powered randomised controlled trials comparing well-matched patient groups with long-term follow-up are required to limit systematic error and enhance external validity. Specific outcome measures should include union, functional assessment, complications and cost–benefit analysis.