Symptomatic malunion after midshaft clavicle fracture in an adolescent patient: a case report of surgical deformity correction using a 3D printed model

Abstract

A 14-year-old boy underwent surgery for symptomatic malunion of the clavicle. This complication, which is uncommon in adults and adolescents, occurred after a displaced midshaft clavicle fracture that had been treated conservatively. Surgery may be considered if functional impairment, pain, weakness, fatigability and neurological symptoms persist in the presence of significant clavicular deformity. Our case was unusual because the patient had a symptomatic malunion and lost range of movement of his shoulder despite a minor degree of clavicular shortening. We adopted an approach used in lower limb deformity correction but not described for the clavicle in which corrective osteotomy was planned and practised using a three-dimensional printed model of the malunited clavicle. A three-dimensional printed model of the mirror image of the opposite clavicle served as a template of normal. Three-dimensional models were printed from the computed tomography data. The patient’s symptoms resolved and he recovered full range of movement and shoulder function following the corrective osteotomy.

Background

Clavicle fractures are the most common long bone fractures in children and comprise 10–15% of all paediatric fractures.¹ Most of these fractures (95%) occur in the diaphysis. The vast majority of paediatric clavicle fractures are treated non-operatively, with union rates reported at 96.3% with complete recovery of shoulder function. Symptomatic malunion was not historically reported as a complication of non-operatively treated fractures. However, a recent study of adolescent clavicle fractures suggests that symptomatic malunion may occur more commonly than previously thought, particularly when the fractures are associated with significant displacement.² Assessing displacement (translation and rotation) in clavicle malunion is difficult. The reliability of assessing clavicle shortening using plain radiographs has been questioned.^3,4 The value of using three-dimensional (3D) models for the assessment and correction of deformity and operative planning is documented in the literature;⁵ however, to our knowledge this has not been described in the context of clavicular malunion and its correction.

This report describes the case of a 14-year-old adolescent with a symptomatic malunion of a diaphyseal clavicle fracture assumed to be due to altered shoulder kinematics. Computed tomography (CT) images were used to understand the translational and rotational deformity. A 3D-printed model was used to plan the corrective osteotomy and he underwent successful surgery.

Case history

Presenting history

A 12-year-old right-hand dominant boy sustained a fracture at the junction of the middle and distal third of the left clavicle while playing rugby. It was a closed injury and there was no associated neurovascular injury. The fracture was angulated approximately 40 degrees, apex superior, and there was less than 25% displacement (Fig 1a). There was no comminution. The fracture was treated conservatively and he used a polysling for three weeks. He was followed up in the paediatric fracture clinic and, at six weeks, the fracture demonstrated healing with callus formation on radiological assessment. At three months, in spite of evidence of radiological union, he continued to experience pain in his left shoulder. His pain was poorly localised and of note the pain was not felt over the clavicle itself. He was unable to elevate his shoulder above the horizontal and his left arm and shoulder were prone to fatigue. He was referred for physiotherapy but his pain and dysfunction persisted and, as a result, he stopped most of his sporting activities and could not return to playing rugby.

Figure 1.

Anteroposterior radiograph demonstrating (a) the initial position of the fracture and (b) the malunited fracture.

He was referred to the shoulder clinic for an opinion 18 months after his original injury because of his persisting symptoms. His general health was good; he had no significant medical conditions. He took no regular medicines and had no known drug allergies.

On examination, he had asymmetrical shoulders with protraction of the left scapula and a visible lump at the site of the fracture. There was trapezius muscle spasm and overactivity and the left shoulder was held higher than the right. There were no signs of hyperlaxity and both the sulcus sign and Gagey’s test were negative. Passive range of movement was full but he was limited to 150 degrees of forward flexion and 90 degrees of abduction actively. With scapular correction, he achieved 170 degrees of forward flexion and 160 degrees of abduction, but with pain at the limits of his range. There was no neurological deficit or evidence of radiculopathy.

X-rays (standard 0-degree anteroposterior view and a 45-degree inferior tilt view of the clavicle) showed a healed angulated fracture of the clavicle (Fig 1b). A CT angiogram excluded compression of the subclavian vessels and confirmed union. A magnetic resonance image showed no evidence of a brachial plexus lesion or shoulder joint pathology.

Preoperative planning

The CT scan showed that there was only 10mm shortening of the left clavicle when compared with the right clavicle. However, we estimated a 16-degree increase in angulation on the axial slices (apex posterior), a 29-degree increase in angulation on the coronal slices (apex superior) and a 32-degree rotational deformity around the long axis of the clavicle.

We printed 3D models of both clavicles with a Lultzbot Taz 5 3D printer (Aleph Objects Inc, Colorado USA), using the manufacturer’s CT protocols, clearly demonstrating the angular and rotational deformities (Fig 2). Mirrored CT images of the right clavicle were used to print a 3D model which served as a reference for a normal left clavicle. A wedge osteotomy was planned to allow restoration of length and correction of both the angular and rotational deformities and then performed on the 3D printed model. An eight-hole ACUMED precontoured left clavicle plate in reverse orientation (medial plate on lateral clavicle and lateral plate on medial clavicle) provided the best fit of the available trial plates to stabilise the osteotomy (Fig 3). It was possible to apply compression across the osteotomy and achieve 50% bone on bone contact (opening wedge with inferior defect). We planned to perform the same procedure using the same implant in theatre and, given the excellent stability of the fixation and good bone-to-bone contact, it was considered that no bone graft would be required.

Figure 2.

The true deformity of the clavicle is easily appreciated with the 3D printed model demonstrating (a) angulation and (b) malrotation.

Figure 3.

Mirrored computed tomography images of the right clavicle were used to print a 3D model which served as a reference for a normal left clavicle and the osteotomy was planned (a), executed and stabilised correcting angulation (b) and rotation (c).

Surgical correction

In consultation with the patient and his parents, the implications and all the potential risks of corrective surgery were discussed. His pain and shoulder dysfunction had persisted for almost two years and the consensus was that he had exhausted conservative options as he had had over six months of appropriate physiotherapy.

The procedure was performed under general anaesthesia with the patient in the beach-chair position. A transverse incision was made along the inferior border of the clavicle. The supraclavicular nerves were protected during the procedure. The deltotrapezial fascia was incised and flaps raised to permit closure of this layer over the plate. Union of the clavicle was confirmed and the malunion was as predicted by the 3D model. The osteotomy was performed and the angular deformity, shortening and malrotation were corrected according to the preoperative plan. Care was taken to avoid damage to the underlying neurovascular structures. At the time of the operation, it was felt that the nine-hole distal clavicular plate provided more secure fixation in the lateral clavicle as the apex of the deformity was quite lateral. The plate was applied achieving compression across the osteotomy by creating an ‘axilla’ and using the compression oval hole of the plate. No bone graft was used. After copious wash with saline and meticulous haemostasis, a standard closure was performed in layers. Bupivacaine 0.5% 20ml with adrenalin 1 : 200,000 was infiltrated for analgesia. The operation was performed as a day case.

Rehabilitation and outcome

The shoulder was protected in a polysling for four weeks. Codman’s pendular exercises and active assisted forward elevation and abduction were started immediately but elevation and abduction were limited to 90 degrees for the first four weeks. No restrictions were placed on range of movement after this. Resistance and strengthening exercises supervised by a physiotherapist were permitted at eight weeks.

Radiological and clinical assessments were performed at regular intervals (Fig 4). The patient was already aware that his shoulder felt ‘more normal’ two weeks after the operation. By 6 weeks, the pain he had preoperatively had completely resolved and he had recovered 160 degrees of forward elevation and abduction. Complete union of the osteotomy and bridging of the inferior defect was evident at 12 weeks. At 12 months, he had no pain, full range of movement and normal shoulder strength and endurance and he had returned to playing rugby without restrictions. In view of the excellent clinical outcome, we could not justify postoperative CT to evaluate correction of rotation and angulation. We confirmed that clavicular length was restored when compared with the other clavicle (less than 1mm side-to-side difference in length).

Figure 4.

(a) Immediate postoperative radiographs (note inferior opening wedge) and (b) radiograph of the left clavicle taken at 12 weeks, demonstrating radiological union across the opening wedge osteotomy.

Discussion

Complications of clavicle fractures in the paediatric population are rare (3.7%);⁶ 100% displaced fractures are 3.2 times more likely to result in complications.⁶ The complication rate in adolescents approaches that seen in the adult population (5–6%). There is an age-related increase in the risk of complication of 18.1% per year.⁶ Strauss et al reviewed 537 paediatric clavicle fractures; 520 were treated non-operatively and, of these, only 13 (2.5%) were associated with complications.⁶ There were six refractures, three delayed unions, three chronic skin irritation due to bony prominence and one non-union.

Malunions are common when clavicle fractures are treated non-operatively but they are rarely symptomatic.⁷ When they occur, symptoms include pain at the malunion site, shoulder pain with overhead activities, weakness, fatigability and neurological symptoms (typically numbness and paraesthesia at the ulnar border of the hand and forearm and occasionally shock-like sensations radiating down the forearm). Much of the literature on symptomatic malunion relates to the adult population. Hill et al found that 31% of patients with shortening of more than 20mm had an unsatisfactory result; these patients complained of pain and local tenderness at the site of malunion but had normal range of movement and shoulder strength.⁸ Eskola et al reported that patients with 15mm or more shortening had significantly more shoulder pain after fracture healing.⁷ Our patient lost active range of movement and this has not been reported in the adult or paediatric literature. Having excluded neurological, glenohumeral and rotator cuff pathology, we concluded that his loss of range of movement was due to the malunion.

Corrective osteotomy has been used successfully to treat symptomatic malunion in the adult population. McKee et al reported the outcome of 15 adult patients who underwent a corrective osteotomy for symptomatic malunion of middle third clavicle fractures.⁹ Patients reported shoulder pain, weakness, fatigability, cosmetic issues and neurological symptoms ascribed to irritation of the infraclavicular brachial plexus. There was no loss of range of movement compared with the contralateral side. The mean shortening was 2.9cm. Postoperatively, there was a significant improvement in shortening (mean 0.4cm) and Disabilities of the Arm, Shoulder and Hand score.⁹ One patient required revision surgery for non-union.

We could identify only one study reporting the use of corrective osteotomy to treat symptomatic malunion of the clavicle in the paediatric population. In their series of 43 paediatric clavicle fractures (mean age 15.4 years), Vander Have et al identified five patients with a mean age of 14.6 years (range 12–18 years) who developed symptomatic malunions.² All five had been treated non-operatively and presented with a variety of symptoms including pain with prolonged overhead activity (n = 2), easy fatigability (n = 1), axillary pain (n = 1) and drooping shoulder with painful bony prominence (n = 1). The authors did not comment on range of movement. Symptoms developed an average of 14 months after the injury (range 6–24 months). The average fracture displacement in this group of patients was 26.6mm (range 20–32mm). Four patients underwent corrective osteotomy and internal fixation. All osteotomies had healed by six weeks and the patients returned to activities at a mean of twelve weeks postoperatively.

The literature is inconsistent regarding the relevance of clavicular shortening to the development of symptomatic malunion.^2,10,11 However, the above studies imply that clavicular shortening is a key consideration and clavicular shortening has been shown to have a negative effect on shoulder kinematics in vivo.¹² There are two accepted methods of assessing shortening: fragment overlap and side-to-side difference. There was no fragment overlap in our case and only 10mm side-to-side difference, which was much less shortening than the apparent threshold for surgery as described in the literature.

Preoperative assessment of CT enabled us to assess angulation and malrotation. We hypothesised that 10mm of clavicular shortening combined with the angular deformity and malrotation of the clavicle that we observed in this case affected scapular posture and shoulder dynamics more than one would expect from the shortening alone, and that correcting this would result in resolution of symptoms and restoration of shoulder movement.

The 3D printed model allowed both clinician and patient to appreciate the true extent of the deformity of the clavicle and was extremely valuable with respect to preoperative discussion with the patient and his family. We found it very useful to practice the planned osteotomy on the model and confirm that correction of length, rotation and angulation could be achieved with an oblique wedge cut and that stable fixation was possible.