A thirty-nine-year-old, right-hand-dominant woman presented seven months after sustaining a displaced right clavicular fracture, which resulted from a fall from a bicycle. She had seen an orthopaedic surgeon one week after the injury; the right arm had been placed in a sling. Because she subsequently became uninsured several weeks after the injury, she had not sought additional care. She presented to us with limited shoulder function and persistent pain at rest and with attempted motion. The medical history was unremarkable, but she smoked half a pack of cigarettes per day. On physical examination, there was gross deformity and tenderness to palpation at the midshaft of the clavicle. Shoulder motion was limited by pain. Neurovascular status was intact. Radiographs of the clavicle demonstrated a widely displaced midshaft clavicular fracture nonunion (Fig. 1). We recommended surgical treatment with ORIF and iliac crest bone graft.
The patient was placed in the supine position with a 1-L bag of saline solution under the ipsilateral scapula and the head in a neutral position to avoid excessive traction on the brachial plexus. An incision following Langer’s lines was centered over the nonunion site. Dissection was carried through the subcutaneous tissues to the site of the nonunion. Fibrous tissue and ectopic bone were excised from the fracture site. Care was taken to stay subperiosteal and to avoid excessive soft-tissue stripping. A microsagittal saw was used to resect 2 mm of bone from the end of each fracture fragment, and a high-speed bur was used to decorticate the intramedullary canal of each fragment. Fracture reduction and fixation was achieved with a lag screw and a neutralization plate. Approximately 5 cc of corticocancellous iliac crest bone graft was placed anterior and posterior to the nonunion site. Satisfactory fracture reduction and plate configuration were confirmed with an image intensifier, and the wound was closed in a layered fashion (Fig. 2).
The patient was evaluated two hours postoperatively and had a 2+ radial pulse as well as a warm, well-perfused extremity, but decreased sensation in the radial, ulnar, and median nerve distributions distally, with intact axillary nerve sensation proximally. Motor function was graded as 5 of 5 in elbow flexion, but there were deficits in other muscle groups: elbow extension was 3 of 5, wrist extension was 4 of 5, wrist flexion was 1 of 5, and hand function (anterior interosseous nerve, posterior interosseous nerve, and ulnar nerve) was 1 of 5. The patient had pain at the surgical site as well as in the shoulder, upper arm, and forearm. There was no evidence of hematoma at the surgical site. The patient’s arm was elevated with pillows to above the level of the heart, and a plan was formed for edema control and serial neurologic examinations. By postoperative day two, both the edema and the patient’s effort with physical examination maneuvers had improved, but the neurologic examination was unchanged. On postoperative day three, a magnetic resonance imaging (MRI) scan demonstrated no sign of a hematoma or other anatomic source that was creating pressure on the brachial plexus. On postoperative day four, a computerized tomography (CT) scan with three-dimensional reconstructions was obtained.
Multiple CT images and reconstructions demonstrated a 2-cm posteroinferior butterfly fragment malunited to the proximal fracture fragment, which was not as apparent on radiography or MRI (Figs. 3-A, 3-B, and 3-C). The CT demonstrated that, after reduction of the fracture fragments, this posteroinferior fragment was compressing the brachial plexus. The decision was made to proceed with surgical excision of the fragment. Intraoperatively, dissection was performed around the medial segment of the clavicle, both posteriorly and anteriorly in a subperiosteal plane. The bone fragment was exposed and excised. The plate was left intact during the excision of the malunited fragment. Neural structures were grossly visible beneath the fragment.
In the immediate postoperative setting, the radicular pain resolved. Sensation was normal in the radial and median nerve distributions and improved in the ulnar nerve distribution. Motor examination was unchanged immediately postoperatively and at discharge two days following the second operation. Electromyography (EMG) at two weeks postoperatively demonstrated a diffuse brachial plexopathy, especially affecting the lower trunk, with all tested nerves in continuity. At approximately four months, a repeat EMG demonstrated a lower trunk brachial plexopathy with evidence of acute muscle denervation with reinnervation by axonal sprouting. On physical examination, manual muscle testing revealed elbow flexion was 5 of 5, elbow extension was 4 of 5, wrist extension was 5 of 5, thumb abduction was 4 of 5, digit extension was 2 of 5, and interosseous strength was 3 of 5. Sensation had improved in all distributions. However, on pinprick testing, sensation was reduced over the dorsal radial aspect of the hand and forearm. At that time, the patient was continuing physical therapy, had a healed clavicular fracture, and was being followed clinically for improvement in neurologic function.
Brachial plexus injury from a clavicular fracture is a rare phenomenon, but one with potentially devastating consequences. There are multiple possible causes of brachial plexopathy from a clavicular fracture. Most reported cases have had a delayed onset of symptoms from the initial injury and have been caused by direct compression on the brachial plexus3-6. Involvement of the medial and posterior cords of the brachial plexus has been reported most frequently7. Acute and direct compression of the brachial plexus by a fragment of clavicular bone has been described3,8,9. Similarly, chronic and direct compression of the brachial plexus by hypertrophic callus, scar, or malunion has been reported4,10-12. Additionally, subclavian artery pseudoaneurysm with corresponding nerve compression has been described after clavicular fracture13. Clinically, patients with pseudoaneurysms present with a pulsatile mass or distal vascular deficits, and the diagnosis is confirmed with a CT angiogram. In our review of the literature, we did not identify any cases of clavicular nonunion with normal neurologic status that developed brachial plexus compression from a bone fragment after ORIF.
Der Tavitian et al. reviewed twenty-four publications and 301 patients treated surgically for clavicular fracture nonunion and found forty-five (15%) reported complications related to soft tissues14. Of the soft-tissue complications, there were two with brachial plexus injury following clavicular nonunion repair15,16. Karaharju et al. reported one patient with moderate signs of brachial plexus compression following repair of a clavicular nonunion; the patient’s motor and sensory function reportedly returned to normal within four months16. Eskola et al. reported one patient who had four different complications: pneumothorax, subclavian vein lesion, air embolism, and transient brachial plexus injury15.
There are multiple potential mechanisms of acute brachial plexus injury following surgical repair of a clavicular nonunion. In our case, the clavicle was effectively shortened during fixation, and, therefore, a traction injury seemed an unlikely cause. Although we extensively debrided the fracture site of all callus and fibrous tissue, we avoided dissection of the medial and inferior clavicle to prevent excessive soft-tissue stripping. There were no preoperative symptoms to suggest the need for more aggressive debridement or neurolysis. On re-exploration, it was clear that, during reduction of the primary fracture fragments, the malunited posteroinferior bone fragment compressed the brachial plexus. Because this is a rare event and a single case report, we cannot recommend use of advanced imaging (CT or MRI) in the preoperative evaluation of every clavicular nonunion in the absence of neurologic or vascular deficit. Intraoperative EMG neural monitoring could have identified brachial plexus compression during the index surgery. However, since we effectively shortened the clavicle and did not conduct an extensive debridement/dissection inferior to the clavicle, we did not consider utilizing intraoperative EMG in this case. We do recommend considering EMG in cases in which the surgical plan involves lengthening the clavicle or extensive dissection or debridement inferior to the clavicle. No specific data exist to guide surgeons regarding a critical amount of lengthening or dissection that would place the neurovascular structures at high risk for injury.
We reported a case of compressive brachial plexopathy following clavicular fracture nonunion repair. The use of CT imaging and three-dimensional reconstruction allowed us to visualize an osseous fragment that was causing compression on the brachial plexus. We believe that these imaging techniques were useful in the initial study of this case of brachial plexopathy after fixation of a clavicular fracture nonunion.