Coracoclavicular fixation techniques for Neer IIb and ‘extra-lateral’ fractures of the distal clavicle: A systematic review

5


Conclusion 30
The present systematic review of coracoclavicular stabilization techniques for unstable Neer IIb 31 and extra-lateral fractures of the distal clavicle demonstrates promising clinical outcomes, 32 including effectiveness and safety. We support the previously proposed modification of the Neer 33 classification to include this unique type of unstable extra-lateral fracture (type IIc) to allow for 34 targeted surgical management. 35 clavicle fractures, whereas they also proposed that type IId can replace type V. Nonetheless, Neer 53 types IIa/IIb and V (or Cho types IIb to IId) are usually unstable and indicative of internal fixation. 6 54 In the literature, there is no optimal surgical technique for managing unstable fractures of 55 the distal clavicle, nonetheless for unique patterns such as IIc. Several surgical techniques have 56 been described and are generally distinguished in two main categories; either rigid internal fixation 57 (T-, locking-, hook-, double-plates, coracoclavicular or acromioclavicular screws, and 58 intramedullary fixation), or flexible osteosynthesis with or without arthroscopic assistance (tension 59 band wiring, sutures and bone anchors, tapes, cortical buttons, and synthetic grafts or allografts), 60 as well as various combinations of these techniques. [7][8][9][10][11] The evidence is even more vague for extra-61 lateral (Cho IIc) fractures, where the lateral clavicle fragment being relatively small, is not always 62 amenable to hold traditional hardware. Levy et al were the first to report on fractures classified as 63 Type IIcout of 48 patients enrolled in the study, 30 were treated with simple coracoclavicular 64 suture stabilization and the rest underwent plate fixation and coracoclavicular augmentationand 65 conclude that there is a need for modifying the original Neer's classification to include type IIc 66 fractures. 12  We exported all captured studies into a reference manager library (EndNote X9) and 102 removed all duplicates. The results were screened by two independent reviewers at two levels: 103 title-abstract and full text screening. We resolved any discrepancies during title-abstract screening 104 stage by including the article by default, and during full text screening by discussion and senior 105 author consensus. 106

107
All relevant data were extracted using piloted forms and exported to a digital spreadsheet demographics, surgical intervention, and outcomes and results. Any discrepancies in the extracted 111 data were resolved by thoroughly inspecting the manuscripts during reviewer meetings. 112

113
Our review did not capture any randomized controlled trials (RCTs), thus the ROBINS-I tool for 114 assessing risk of bias in non-randomized studies was used. 15 We stratified the risk for confounding, 115 selection of participants into the study, classification of interventions, deviations from intended 116 interventions, missing data, measurement of outcomes, selection of the reported result and overall 117 bias. Overall risk of bias was considered low risk if all domains were determined at low risk; 118 moderate risk if at least one of the domains was determined at moderate risk but none as serious; 119 serious risk if at least one of the domains was determined at serious risk but none as critical; and 120 critical risk if at least one of the domains was determined at critical risk. 121 122 We synthesized all studies qualitatively using descriptive statistics, where applicable. We 123 categorized the reported surgical techniques in 2 types: arthroscopic assisted coracoclavicular 124 stabilization with buttons, with or without interfragmentary sutures or tension band, and open 125 coracoclavicular stabilization with buttons, subcoracoid sutures, mersilene tapes, cables or suture 126 anchors. We reported findings for two main outcomes: effectiveness and safety. Effectiveness was 127 evaluated by synthesizing and summarizing data from objective clinical scores as reported by the 128 included studies. Safety was assessed by calculating the frequency of complicationswe 129 categorized complications into major (new fracture, implant failure, nonunion, coracoid fracture, 130

Data synthesis and analysis
Our database search following duplicate removal yielded a total of 564 records. We retrieved the 136 full texts of 98 manuscripts to screen in their entity, whereas 466 records were excluded ( Figure  137 1). Only 21 were deemed appropriate for inclusion in our qualitative synthesis ( Table 1). The most 138 common reason for exclusion was studies reporting data on management methods other than 139 coracoclavicular surgical fixation techniques (n = 49) ( Figure 1). 140

Characteristics of included studies 141
The total number of reported Neer IIb/Cho IIc distal clavicle fractures managed with a 142 coracoclavicular stabilization technique in all studies was 421, ranging from 6 to 45 patients. In  Table 1.  Table 1). One arthroscopic study utilized All studies were retrospective with a high or moderate ROBINS-I overall risk of bias assessment. 165 The heterogeneity of the operative techniques and the different evaluation methods prohibited us 166 from performing a meta-analysis. Thus, our systematic review has a low level of evidence (IV). 167

Effectiveness 168
Clinical results were reported with various clinical scores, including the Constant score (10 169 studies), ASES (5 studies), UCLA (6 studies), Oxford score (2 studies), DASH (4 studies), 170 Karlsson's criteria (1 study) and Modified Shoulder Rating Scale for Clavicle Fractures (1 study) 171 ( The overall major complication rate was 2.6% (8 nonunions, 1 coracoid fracture, 1 fracture 179 between clavicular holes and 1 cable breakage), whereas minor complications were present in 54 180 out of 421 patients (12.8%) ( Table 2) found that the complication rate was significantly higher with the use of the hook plate (40.7%) or 211 tension band wiring (20.0%), compared to coracoclavicular (4.8%), intramedullary (2.4%) and 212 interfragmentary fixation (6.3%). 11 A limitation of this study was the uncertainty of whether type 213 II fractures were constituted of mainly type IIa or IIb patterns, as most included studies did not 214 specify the type II subclassification. 215 Special attention is required for some specific complications following coracoclavicular 216 button fixation, including migration, slippage, coracoid or clavicle erosion, and button subsidence 217 though the tunnels. This can lead to re-dislocation of the proximal clavicle and loss of fracture 218 reduction. A similar problem has been encountered with the use of these techniques in 219 acromioclavicular joint dislocations. 30 The aforementioned complications may be attributed to the 220 mispositioning of the tunnels and the excessive tension of the paired button bearings, which increase the force of slippage, especially when laid on the uneven face of the clavicle or the 222 coracoid process. Furthermore, when the tension caused by the button on the clavicle or the 223 coracoid process is too concentrated, this can lead to bone erosion. This phenomenon is usually 224 encountered when a broad tunnel receives a button with a relatively small total area. Dog-ears, flat 225 buttons and two-tail systems (triple buttons) can solve this problem by distributing the tension 226 forces evenly. 31, 32 227 We find that precise interpretation of distal clavicle fracture patterns can guide the surgeon 228 in choosing the most appropriate surgical technique, however, interpretation of unstable types (IIa, 229 IIb, IIc and V) is inherently challenging. 3-5, 10 Neer's classification has been widely used over the 230 past decades to classify these fractures, however, according to Bishop et al the interrater agreement 231 is only fair for distal fragment size and type, moderate for stability and treatment approach, and 232 slight for type IIb fractures. 33    J o u r n a l P r e -p r o o f