Clavicular bone defects managed with free vascularised fibular grafting: evidence to date

Reconstructive surgery of the clavicle using free vascularised fibula grafting (FVFG) is sometimes required for the management of severe bone loss or non-union. As the procedure is relatively rare, there is no universal agreement on the management and outcome. This systematic review aimed to first, identify the conditions for which FVFG has been applied; second, to gain an understanding of the surgical techniques used; and third, to report outcomes related to bone union, infection eradication, function and complications. A PRISMA strategy was used. Medline, Cochrane Central Register of Controlled Trials, Scopus and EMBASE library databases were interrogated using pre-defined MeSH terms and Boolean operators. Quality of evidence was evaluated based on OCEBM and GRADE systems. Fourteen studies based on 37 patients were identified with a mean follow-up time of 33.3 months. The most common reasons for the procedure were: fracture non-union; tumours requiring resection; post-radiation treatment osteonecrosis and osteomyelitis. The operation approaches were similar, involving graft retrieval, insertion and fixation and vessels chosen for reattachment. The mean clavicular bone defect size was 6.6 cm (± 1.5), prior to FVFG. Bone union occurred in 94.6% with good functional outcomes. Complete infection eradication occurred in those with preceding osteomyelitis. The main complications were broken metalwork, delayed union/non-union and fibular leg paraesthesia (n = 20). The mean re-operation number was 1.6 (range 0–5.0). The study demonstrates that FVFG is well tolerated and has a high success rate. However, patients should be advised about complication development and re-intervention requirement. Interestingly, overall data is sparse with no large cohort groups or randomised trials.


Introduction
The clavicle is a double curved s-shaped long bone articulating with the sternum medially and acromion laterally, and is stabilised by strong ligaments at either end [1].Maintaining the integrity of the clavicle is important for normal shoulder function, the avoidance of compression of underlying structures (e.g., brachial plexus and axillary artery) and for optimum respiratory function [2].
Injury and dysfunction of the clavicle most commonly follows a fracture occurring as a consequence of direct trauma to the shoulder [3,4].Clavicle fractures are fairly common and account for approximately 2.6-4% of all fractures [5,6].In many of these cases, management is nonsurgical, but when surgery is required, an open reduction and internal fixation (ORIF) with a plate is most frequently used in the first stages to stabilise the bone and encourage healing [7,8].In the UK, surgical intervention for clavicular fractures ranges from 2 to 4% of all fractures [9].
If unsuccessful, the plate is removed and another inserted with an iliac crest autologous bone graft, for enhancement of the fracture healing response.On failing this, reconstructive surgery is then considered.A similar scenario may follow other clavicular disorders, such as tumour infiltration, osteomyelitis or bone necrosis secondary to radiotherapy [10][11][12], although in these latter cases, reconstructive surgery may be required as an earlier option.
Taylor et al., described a reconstructive approach known as the free vascularised fibula graft (FVFG) to manage significant long bone defects [13].The principles of this technique relate to the abundant vascularity of the fibula, its similarity in bony shape to the clavicle, and its functional adaptation to the recipient site [14].However, as the procedure is relatively uncommon [15], there is a paucity of literature specifically describing the outcomes.
The objectives of this review were to: first, identify the conditions for which FVFG has been applied; second, to gain an understanding of the types and choice of surgical techniques used; and third, to report outcomes related to bone union, infection eradication, functional and complications.

Search strategy and criteria
The protocol for this systematic review was based on the preferred reporting items for systematic reviews and metaanalyses (PRISMA) guidelines [16], and was created prior to data extraction.A list of Medical Subject Headings (MeSH) terms and Boolean operators were compiled: (clavicle OR clavicular) AND (non-union OR pseudarthrosis) AND (management OR free vascularised fibular graft).These words were utilised to search Medline (through the PubMed search engine), Cochrane Central Register of Controlled Trials (CENTRAL), Scopus and EMBASE databases.The search strategies implemented are detailed in Appendix 1.

Study selection
Inclusion criteria were established using the population intervention comparison outcomes (PICO) approach [17]: Population: adults (over 18) with clavicular non-union and concomitant osseous defects.There were no limits on sex, ethnicity or co-morbidities of the individuals included.Intervention: FVFG.Comparator: management strategies used to treat clavicular non-union e.g., ORIF or bone grafting, excluding FVFG.Outcomes: the primary outcome measured was bone healing.Secondary outcomes included infection eradication; functional outcomes; and complications including unplanned re-operations.Exclusion criteria included: reviews, editorials and viewpoints, subjects aged 16 years or less, congenital cases of clavicular non-union and cases in which a complete neo-clavicle was required.
All studies were considered for eligibility, with no restrictions on publication date or language applied.Titles and abstracts were screened for relevance prior to full inspection.The reference lists of all eligible studies were reviewed to isolate any articles that may have been missed in the initial database search.Duplicate articles between the databases were removed and the full texts of all studies meeting the inclusion criteria were obtained.To increase the reliability of data extraction, two reviewers blindly performed the study selection and data extraction.Any disagreements between reviewers were resolved with discussion with the senior author.

Data collection
Data was extracted and collated using a purpose-designed Microsoft Excel spreadsheet.The following data were recorded: (1)

Assessment of methodological quality
Methodological quality of the included studies was assessed and graded using the OCEBM 'Levels of Evidence' guidelines [18].
The overall quality of evidence in this systematic review was evaluated using the grading of recommendations, assessment, development and evaluation (GRADE) system [19].Recommendations were classified as either high, moderate, low, or very low according to the authors' interpretation of the true effect in the study compared to the estimated effect.This approach involved grading the evidence included based on the criteria for risk of bias, imprecision, inconsistency, indirectness and publication bias.

Statistical analysis
Descriptive statistical analysis (e.g., mean ± standard deviation (SD), mean ranges, ratios and percentages) was collated and reported in this study.

Search results
The PRISMA flowchart is shown in Fig. 1.
102 Medline articles, 11 CENTRAL articles, 154 Scopus articles and 121 EMBASE articles were obtained.In addition, a further search of records yielded an additional 4 studies; this provided a baseline of 392 studies in total.
In addition, Appendix 2 provides an overview of the GRADE analysis assessment, which demonstrated the quality of evidence to be Low for all analyses, as there were no large observational studies or randomised controlled trials on this subject.

Patient demographics and baseline characteristics
Of the 37 individuals reported in this review with a mean patient age of 44.8 (± 12.8) years old, 19 were men and 16 women; the sex of 2 cases was unreported (1.2:1 male to female ratio) (Table 1).

Initial clavicular injury and surgical procedures prior to FVFG intervention
The causes of clavicular bone defect have been classified into four main categories: fracture non-union, tumours, osteoradionecrosis following radiation therapy for tumour treatment and osteomyelitis.
The mean clavicular bone defect size was 6.6 cm (± 1.5), prior to vascularised fibular grafting.The mean size of these defects in the sample ranged from 4.5 cm to 9.5 cm (Table 1).

Discussion
The treatment of large bone defects remains a challenge for reconstructive surgeons.Usually, this group of patients has been through a prolonged clinical journey having undergone several operations that have failed to address the original problem, usually being fracture non-union and/or chronic osteomyelitis.The need to resect the avascular, dead bone leads to bone loss and the development of bone defects.
Evidence was sought for the use of a vascularised fibular bone graft for the management of significant clavicular bone defects requiring reconstructive surgery.We examined the aetiology, patient characteristics, the variations in surgical technique and outcomes.
Overall, there was a slightly higher predisposition towards males requiring this surgery.The mean age was 44.8 years which may reflect a higher rate of falls, an increasing risk of bone pathologies or less preponderance to healing when compared to a younger age group [4].Direct trauma accounted for approximately 80% (30/37) of the cases as the primary mechanism of injury.In these cases, fibular graft surgery often followed several preceding operative interventions (mean previous operations 2.5), such as an ORIF with plating, bone grafting (often from the iliac crest) and debridement.The remaining cases mostly related to infiltrative and/or destructive disease processes, such as that caused by tumour, radiotherapy or infection; in these situations, the procedure was more likely to be done as an earlier intervention [34,35].The mean cortical defect size across studies was 6.6 cm (range 4.5-9.5 cm); their relatively large size reflecting the need for a more aggressive approach.In some cases, the large size of the defect reflected the consequences of multiple previous operative resections and debridements of the bone ends from previous attempts at surgery or because the lesion was primarily large e.g.due to neoplasm.
Although the general principles of harvesting and implanting a graft was similar between studies, there were variations in the operation technique.The choice of a specific technique might be dependent on the degree of bone loss, underlying reasons for the bone loss as well as the personal preferences of the surgeon.
The reported overall outcome of a fibular graft was 94.6% (35/37) for a successful union.Eleven different functional outcome measures were employed across the studies highlighting a lack of standardisation.The reported complications related to either the clavicle itself (21 patients) or the fibular donor site (15 patients).With respect to the clavicle, most complications related to the metalwork and vascular tree.In contrast, paraesthesia and weakness were the main consequences of the fibular procedure [30].The mean number of re-operations was 1.6, with a range of 0 to 5.0 additional procedures.It is therefore pertinent for surgeons to warn patients of an increased risk of re-interventions and chronic pain/ paraesthesia to the donor or recipient site.
The main limitation of the study is the relatively small sample size which has prevented formal statistical analysis.It is also unknown how common this procedure is performed within practice and whether only cases with a positive surgical outcome are published.It was also noted that there were differences in what data was reported within studies resulting in variability in some demographic and outcome data.The two cases that failed to unite did not provide data on co-morbidities or previous surgical procedures [20,28].
The strengths of the paper relate to it being a systematic review which followed a structured strategy for data collection and analysis.Integrity of the data was optimised by discussions of the articles between authors.During preparation of this manuscript, another systematic review on the same topic was published [33].This was based on data up to January 2020 from only 3 search engines (including Google scholar) and included contrary to our strategy, data on paediatric and congenital cases [33].Despite this, we feel that our study provides further information to the subject area for reconstruction of clavicular defects in adult patients with similar aetiopathogeneses related to fracture non-union and chronic osteomyelitis.
In summary, this study has highlighted that the use of a FVFG, when applied in specific situations, often has successful clinical, functional and radiological outcomes.However, the lack of standardisation of procedures and outcome measurement, and the available small number of patients reported makes it challenging to provide a comprehensive evaluation of the technique.Further studies with larger patient sample sizes are desirable to provide more robust evidence and facilitate a meta-analysis in this field.

Table 1
Summary of study, patient, and initial injury characteristics

Table 2
Characteristics of surgical procedure and outcome measures