a Preoperative radiograph of distal clavicle nonunion. b Post-operative radiograph of distal clavicle nonunion following open reduction and internal fixation with a precontoured distal clavicle plate and a lag screw
Recent studies of displaced midshaft clavicle fractures have shed new light on the reported incidence of nonunion. In 1997, Hill et al. reviewed 52 patients with displaced midshaft clavicle fractures treated nonoperatively . They reported a nonunion rate of 15% (8 out of 52) and an unsatisfactory clinical outcome in 31% of patients. In a review of 2144 displaced midshaft fractures collected from the literature between 1975 and 2005, Zlowodzki et al. reported a nonunion rate of 15.1% (Table 3.1) . This is markedly higher than the rate described by Neer and Rowe. Several other recent studies have reinforced these findings [8, 17, 23–27]. In 2007, a study conducted by the Canadian Orthopedic Trauma Society compared nonoperative treatment versus plate fixation for displaced midshaft clavicle fractures . This series found that there was a significantly lower rate of nonunion in the operative group (62 patients, 2 nonunions, rate of nonunion 3%) compared with the patients treated conservatively (49 patients, 7 nonunions, rate of nonunion 14%). These findings are helpful in establishing that primary fixation may benefit young and active individuals with fully displaced midshaft clavicle fractures. An assessment of a patient’s injuries as well as their functional expectations remains one of the most important considerations regarding decisions for treatment . While operative treatment is typically effective for established nonunions, primary prevention of nonunion would be preferential (Figs. 3.2 and 3.3).
Nonunion of the clavicle following various treatments. Meta-analysis of nonoperative treatment, intramedullary pinning, and plate fixation of displaced midshaft fractures of the clavicle from series published in 1975 through 2005 (From Zlowodzki et al. , with permission)
(n = 159)
Plating (n = 460)
Intramedullary pinning (n = 152)
Total (N = 771)
a Preoperative radiograph of midshaft clavicle nonunion in an active 28-year-old man fourteen months after fracture. b Post-operative radiograph demonstrating union of midshaft clavicle nonunion following open reduction and internal fixation with a precontoured clavicle plate. Note the solid “bridge” portion of the plate, providing extra strength at the nonunion site
a Preoperative radiograph of midshaft clavicle nonunion in a 39-year-old man who had failed two prior attempts at operative fixation. There is significant bony defect. b Post-operative radiograph of midshaft clavicle nonunion following open reduction and internal fixation with a precontoured plate and an intercalary tricortical iliac crest bone graft
Several recent studies have indicated that certain patient populations pose a greater risk of nonunion, malunion, and/or poor functional outcomes [8, 23, 25, 28]. Reported risk factors for nonunion include: a “refracture,” clavicle shortening (>15–20 mm), female gender, fracture comminution, increasing fracture displacement, older age, severe initial trauma, and unstable lateral fracture (Neer type II) [2, 21, 28–32]. It is highly likely that there are other factors that increase risk of nonunion that requires further clarification, such as multiple ipsilateral rib fractures or associated scapular/glenoid fractures [14, 32–34].
With the Allman classification from 1967 , clavicle fractures can be divided into three groups; Group 1, fractures of the middle third; Group 2, fractures of the lateral third; and Group III, fractures of the medial third. These three groups can be further subdivided into nondisplaced and displaced fractures. Within Group 1, there is a further subgroup that consists of fracture comminution and one or more displaced intermediary fragments. In 1998, Robinson presented a new and more comprehensive classification for clavicle fractures . This was performed following a study of 1000 consecutive patients who were treated for isolated clavicle fractures in the Orthopedic Trauma Unit at the Royal Infirmary of Edinburgh. The age, gender, and mechanism of injury were recorded for each patient within 72 h of the injury. The anatomical site, configuration, type, and extent of comminution were also recorded for each fracture. Radiographs from twenty randomly selected fractures aided in the development of this new classification system.
The findings of Robinson’s epidemiological study estimated that in patients over 13 years of age, clavicle fractures occurred at a rate of 29.14 per 100,000 annually . Prior to this, Nordqvist and Peterson had found the incidence of clavicle fractures to be 64 per 100,000 annually . The mean age for clavicular fractures was 29 years in men and 45 years in women. In that series, clavicle fractures occurred most frequently in males under the age of 30 years as a result of a sports injury or a road traffic accident (RTA). The majority of fractures (89%) healed without complication when treated conservatively. Nonunion was only seen in one nondisplaced Type A fracture (Type 3A-1). The rate of delayed union and nonunion was far higher in the displaced Type B fracture cohort, with a rate of 2.7 and 4.8%, respectively. They occurred almost entirely following a displaced diaphyseal (type 2B) or displaced lateral end (type 3B) clavicular fracture. The odds ratio (OR) for delayed union or nonunion following a type 2B fracture compared with a type 2A fracture was 18 (meaning that a patient with a 2B fractures is eighteen times more likely to develop a delayed or nonunion than a patient with a 2A fracture pattern). The OR for delayed union or nonunion following a type 3B fracture compared with a type 3A fracture was seventy-five. However, due to the increased prevalence of type 2B fractures, more patients from this cohort were seen with a delayed or nonunion. High-energy injuries such as falls from a height, RTA, and direct violence indicated a greater incidence of delayed and nonunion compared with low-energy mechanisms of injury. Differences in the rate of union were not significant amongst the age or gender cohorts.
3.4 Rates of Nonunion in Displaced Clavicular Fractures
Nonunion is significantly more common in displaced clavicular fractures, particularly of the midshaft . A direct correlation has been demonstrated between increased degrees of displacement and poor functional outcomes . The study conducted by Robinson et al.  demonstrated that patients with a displaced, comminuted midshaft clavicular fracture had a nonunion rate of 21%.
Neer and Rowe initially reported low rates of nonunion [2, 4]. Subsequent studies have shown significantly higher rates of nonunion. In 1986, Eskola et al. reported a nonunion rate of 3% . In 1997, Hill et al. reported an exponentially higher nonunion rate of 15% . In 2004, Robinson et al. found an overall nonunion rate of 6.2: 8.3% of medial fractures, 4.5% of diaphyseal fractures, and 11.5% of fractures to the lateral end . Similar to Hill, Zlowodzki et al. reported a nonunion rate of 15.1% in 2005 . The Canadian Orthopedic Trauma Society’s 2007 trial reported a rate of nonunion in the operative group of 3% compared with 14% in the nonoperative group . Most recently in 2011, Kulshrestha et al.  conducted a study of 73 patients, aged 20–50 years, who were allocated to operative or nonoperative treatment for their displaced midshaft clavicle fractures. In the operative group, none of the 45 patients went on to nonunion. In the nonoperative group, eight patients out of twenty-eight developed nonunion which resulted in an extremely high rate of nonunion (29%) . New knowledge regarding the potential rate of nonunion for this injury has come to influence management options for orthopedic surgeons who treat these fractures.
3.5 Possible Risk Factors for Nonunion
Murray et al.  sought to determine risk factors for nonunion in an adult population of patients with displaced midshaft clavicle fractures. They retrospectively reviewed 941 patients who were at least 18 years of age, who had initially received nonoperative management for their injuries from January 1994 to December 2007. No nonunions were reported in the 184 patients who were under 18 years of age, and thus, they were excluded from the study. The risk of nonunion in children was zero, and age was not a significant risk factor for nonunion in the adult population. The study population was typical of most midshaft clavicular fracture populations in that it consisted mainly of young and active men. One hundred and twenty-five of these patients went on to develop nonunion which constitutes a risk of nonunion of 13.3% (95% confidence interval [CI], 11.3–15.6%). Using bivariate analysis, the significant factors for a higher nonunion rate were female gender, smoking, increased fracture displacement, overlap, translation, and comminution. In the multivariate analysis, the significant factors were smoking (OR, 3.76), comminution (OR, 1.75), and increased fracture displacement (OR, 1.17) (Fig. 3.4). The negative effect of smoking on fracture healing has been shown in previous studies [40–42]. This is the first time that smoking was identified as a risk factor for clavicular nonunion since earlier studies that investigated whether smoking was a risk factor showed no correlation [21, 28]. However, poorer outcomes in patients with comminution had been shown in prior studies [8, 22, 27, 40]. Murray et al. also identified comorbidities that could potentially increase the risk of nonunion which included rheumatoid disease, immunocompromise, renal failure, epilepsy, and use of drugs such as corticosteroids and those interfering with vitamin D metabolism . Since many patients with a high risk of nonunion will not develop this complication, and many patients with fewer risk factors do go on to nonunion, these risk factors are best used as a guide for surgeons when making a plan for management of this injury. This knowledge would ideally improve the functional outcomes for certain patients who would benefit from surgical intervention, while simultaneously avoiding unwarranted surgery for others .
a Preoperative radiograph of a midshaft clavicle fracture in a 25-year-old female. b The patient was treated nonoperatively and went on to develop a symptomatic atrophic nonunion with severe displacement. c Post-operative radiograph of midshaft clavicle nonunion following open reduction and internal fixation with a precontoured plate and an iliac crest bone graft. d Radiograph of healed midshaft clavicle fracture at final follow-up
3.6 Treatment Options
Nonunion is generally defined as the lack of radiographic healing at six months post-initial injury and delayed unions are typically defined when there is progression of healing, but the fracture has not achieved radiographic union at three months . Symptomatic clavicular nonunion is typically painful and debilitating for active patients, causing local pain, persistent deformity, and shoulder weakness resulting in neurologic symptoms consistent with brachial plexus impingement. Treatment options vary from nonoperative techniques that include symptomatic treatment with a sling or figure-of-eight bandage, to noninvasive techniques involving electrical stimulation or low-intensity ultrasound [5, 7, 15, 43], and surgical intervention. Asymptomatic patients with a radiographic nonunion who have full clinical function do not require surgical intervention and respond well to conservative treatment.
Clavicle fractures have traditionally been managed nonoperatively due to studies published in the 1960s that indicated low nonunion rates [2, 4]. This treatment involves mobilizing the shoulder in a sling or figure-of-eight bandage. In a randomized, controlled trial that compared the treatment of clavicle fractures with a sling versus a figure-of-eight bandage, the functional and cosmetic outcomes were found to be identical between the two groups . Recent studies have shown that the risk of nonunion following nonoperative treatment of midshaft clavicle fractures ranges from five to twenty per cent [8, 18, 21, 22]. Patients with displaced fractures are at an even greater risk of developing nonunion [8, 21, 22]. Patients who undergo primary fixation of their displaced midshaft clavicle fractures have better functional outcomes than those treated nonoperatively . Additionally, patients who undergo secondary fixation due to the development of nonunion following nonoperative treatment have results that are somewhat inferior to outcome following primary fixation. [18, 45]. This would suggest that primary fixation for displaced midshaft clavicle fractures may be the preferred course of management for certain patients . However, as Murray et al.  and McKee  suggest, it is important to identify a patient’s risk factors for nonunion in order to assess their suitability for surgical treatment to optimize management.
Methods of intervention include excision [37, 46–49], intramedullary fixation [14, 50–52], and open reduction and internal fixation (ORIF) with cortical  or autogenous bone graft [32, 37, 40, 54–64]. Dual plating with 100% union rates has also been reported . Recent studies have shown that young and active patients with symptomatic nonunion benefit from surgical treatment. However, residual functional impairment can still occur . The series presented by Robinson et al.  found that the highest rate of nonunion occurred in elderly female patients with diaphyseal clavicle fractures. However, due to the reduced functional demands of elderly patients, it is possible that surgery may not be required in these individuals. This is likely not the case for active and young patients who still account for many of the nonunions that occur in clavicle fracture cases. Poor function can hinder lifestyle, and for this reason, surgical intervention is usually desirable since it has been shown to significantly improve long-term shoulder function .
Clavicle nonunions are primarily treated with reconstructive procedures, and in rare cases, a salvage procedure may occur when there are few or no options left for a patient . Salvage procedures have involved partial or total resection of the clavicle bone (claviculectomy) as well as excision of a bony prominence in order to relieve local pain caused by symptomatic nonunion.
3.8.1 Reconstructive Procedures
The primary surgical treatment of clavicle nonunions involves reconstructive procedures. There are a number of methods used in the treatment of clavicle nonunions, with an increasing body of literature to describe such methods and cases (Table 3.2) [14, 37, 40, 50, 54–58, 60, 62, 68, 69]. There are two main operative treatment options used to achieve clavicular union: plate fixation and intrameduallary screw/pin fixation. Plates are typically fixated to the superior surface of the clavicle, but good outcomes have also been reported with compression plates fixated to the anteroinferior surface .
Treatment of clavicular nonunion with various methods of operative intervention
No. of patients who achieved union
Age (y) (min)
Age (y) (max)
Ballmer et al.  (mean follow-up 8.6 y)
35/37 (16 patients had primary operative treatment)
Decortification with plate osteosynthesis. Autogenous cancellous bone graft (24 cases), tricortical, iliac crest intercalary graft (9)
2 persistent symptomatic nonunions
Bradbury et al. 
(mean follow-up 7 y post-surgery)
21 | 10
+1 true pseudoarthrosis
17 AO reconstruction plate with autologous cancellous bone graft, 15 AO dynamic compression plate with autologous cancellous bone graft
Boyer and Axelrod 
Excision with 3.5-mm pelvic construction plate or dynamic compression plate, lag screw for interfragmentary compression, cancellous bone graft
Davids et al. 
AO reconstruction plate with iliac crest bone graft
1 patient had refracture 1 year post-hardware removal
6 (1 year post-surgery)
Ebraheim et al. 
(mean follow-up 12.9 mo)
ORIF with reconstruction or dynamic plate, autogenous bone graft in 14 cases, double plating in 3 cases
2 cases persistent mild pain, 1 hardware failure (healed post-revision), 1 hardware failure (persistent nonunion, pain subsided)
Endrizzi et al. 
42/47 (1 patient LTF, 1 deceased)
ORIF-27 curved pelvic reconstruction plates, 16 straight pelvic reconstruction plates, 4 straight dynamic compression plates
14 cases demineralized bone matrix, 3 autogenous iliac crest graft, 1 resected rib graft
3 revision surgeries due to implant loosening
Eskola et al. 
+2 patients treated with resection
Twenty-one cancellous bone grafts were executed, with eighteen of these involving rigid plate fixations and one Kirschner pin fixation. In two cases, bone grafting was used solely. In one case, only plate fixation was performed
4 cases of clavicle shortening
Kabak et al. 
ORIF with dynamic compression plate (DCP) or low-contact (LC-DCP) and autogenous corticancellous chips or sculptured graft
1 incomplete brachial plexus palsy, 2 revision surgeries went on to union
Laursen and Dossing 
(mean follow-up 24 mo)
Compression plate fixation with autologous cancellous bone graft
1 revision due to screw loosening, 1 revision due to fracture of plate
Manske and Szabo 
Compression plate and iliac crest bone graft
Olsen et al. 
1 failed osteosynthesis
ORIF with 3.5-mm plate and autologous iliac crest bone graft
1 persistent nonunion, 2 patients “fair” outcome
In patient with persistent nonunion, hardware removed due to loosening of screws
Capicotto et al.  (mean follow-up 4 y)
Steinman pin fixation and only iliac crest bone graft
Enneking et al.
(mean follow-up 40 mo)
ORIF with Rushpin fixation and iliac crest bone graft
None (aside from 1 case of nonunion)
3.9 Open Reduction Internal Fixation with Autogenous Bone Graft
3.9.1 Surgical Method
ORIF with a compression plate and iliac crest bone graft is considered the gold standard for treatment of clavicle nonunions [1, 7, 71]. Due to the clavicle’s close proximity to the subclavian vascular bundle and brachial plexus, surgical technique must be precise so that neurovascular injury is avoided . This surgical process involves carefully dividing and preserving the myofascial layer and identifying the superior surface of the clavicle. Once this is accomplished, the two ends of the nonunion must be identified and mobilized. Derotation of the (usually anteriorly rotated) distal fragment typically allows the superior surface to be exposed which enables plate fixation on the flat superior surface. The proximal and distal ends of the nonunion are then reduced. If there is excess callous on the superior surface, then it should be ronguered away to create a flat superior surface that facilitates placement of the plate on the bone. This excess callous should be saved, morcellized, and later inserted into the nonunion site. The use of a lag screw or small K-wire to hold the reduction while a plate is applied to the surface can be very beneficial . Due to the complexity of the bone’s structure and the multidirectional biomechanical forces that act on the nonunion, at least three screws should be placed on both sides of the fracture site to stabilize the bone [57, 72]. In a series in which clavicular nonunion were treated with short plates (4 hole semi-tubular), there was a high risk of failure reported . In hypertrophic nonunions, the residual autograft from the local bone should be applied to the nonunion site followed by a standard closure. In cases of atrophic nonunion, an autograft from the iliac crest should be applied to the fracture site (see Fig. 3.2) . If there is significant bone loss with shortening, then it has been proposed that an intercalary graft be used (Fig. 3.3) [7, 59]. The goal in treating these fractures is to restore the length equal to the uninjured contralateral size. Preoperative radiographic and clinical evaluations should be conducted to determine the length of the uninjured side. In the instance of clavicle shortening, it is generally accepted if the bone is shortened by ≤1 cm. If there is significant bone loss with shortening, then it has been proposed that an intercalary graft such as a tricortical autogenous iliac crest bone graft from is used from the contralateral hip. Previously, it has been suggested that low-contact dynamic compression plates are a superior choice for fixation of clavicular nonunion due to its increased ability to be contoured and preserve blood supply to the underlying bone fragments due the plate’s structured undersurface . There have been numerous reports of favourable outcomes and high rates of success (up to 100%) with this method of intervention [54, 58, 60, 69, 74, 75]. It is important that surgeons make certain that an implant of the correct size and length is used in order to aid in successful union of the fracture. The availability of precontoured plates designed specifically for the clavicle, which have been shown to be biomechanically equivalent to compression plates, makes them the implant of choice in this setting [7, 30].