Fig. 4.1
An 82-year-old female two years status-post open reduction and internal fixation with standard T-plate with wound drainage, hardware failure, and surgical neck nonunion on anteroposterior (AP) (a) and lateral (b) views. Underwent stage one of a planned two-stage procedure with removal of hardware, irrigation, debridement, and placement of antibiotic cement spacer, seen one month postoperatively on AP (c) and lateral (d) views. Patient is now six months post-surgery, comfortable and able to perform activities of daily living with motion up to shoulder-level. She has declined her second stage in favor of no further surgery
The surgical indications for nonunions of the proximal humerus include debilitating pain and functional deficits. Some smaller studies found that up to 50% of patients with proximal humerus nonunions are minimally symptomatic and quite functional [19, 33]. A case report study that reviewed the complications of locking plates used for the treatment of proximal humerus fractures found that although the Constant scores were significantly lower for patients who developed nonunions (45 vs. 68), only two of the four patients with a nonunion opted for revision surgeries [34]. A thorough conversation should be undertaken with patients in order to assess their pain and functional deficits and educate them about the risks of revision surgery so the patient may make an informed decision.
4.7.2 Surgical Treatment
Many techniques have been described for the repair of proximal humeral surgical neck nonunions. Plating techniques using conventional T-plates, fixed angle devices such as blade plates or anatomically designed locking plates have been described. Intramedullary implants have also been utilized, ranging from older implants such as Rush rods or Enders nails in conjunction with tension banding, to modern rigid interlocked intramedullary nails. In the setting of bone loss or atrophic nonunion s augmentation with allograft or autograft bone is necessary and structural reinforcement with fibular strut grafts has been found to be biomechanically advantageous. A successful nonunion repair requires adequate bone quality and a proximal fragment that is large enough to obtain purchase in. Although fixed angle devices and locking screw technology improve biomechanical stability in the setting of osteoporotic bone, the size and quality of the fragments should be carefully assessed before osteosynthesis is performed. Significant medial calcar comminution has also been described as a harbinger of difficulty establishing a stable configuration that must be addressed when undertaking fixation. Several techniques have been described, including impaling the head onto the shaft, reconstructing the calcar, using calcar screws to substitute for the medial cortex, or using an intramedullary fibular strut graft [8].
4.7.3 Avascular Necrosis
Humeral head viability is a consideration when deciding on the operative treatment of acute fractures since osteonecrosis and the resultant humeral head collapse may lead to poor results. However, the incidence of avascular necrosis and its functional implications are not fully understood. A review of proximal humerus fractures treated non-operatively found a 2% rate of AVN (12 studies and 650 patients) [11]. Similar reviews of fractures treated using locked plating constructs found rates of AVN ranging from 7.9 to 10.8% [12, 35], although only four of the 51 patients who developed AVN chose to undergo further surgical treatment [35]. Patients who develop AVN after treatment of acute proximal humerus fractures have lower Constant scores (average 46) [36] and a decreased likelihood of achieving good or excellent results when compared to those who do not develop AVN [37].
Gerber et al. performed a subgroup analysis of patients with AVN. They found even with AVN and collapse, Constant scores of patients with an anatomical reduction were similar to patients who had undergone a primary hemiarthroplasty for a proximal humerus fracture. This suggests that patients treated with osteosynthesis after an appropriate reduction who develop AVN may function as well as patients who undergo shoulder hemiarthroplasty [36]. Although the development of AVN adversely affects Constant scores, the functional limitations that result do not drive patients to undergo further surgical intervention at high rates, and they function at levels similar to patients whose fractures were treated using a hemiarthroplasty acutely. These studies are of patients with acute proximal humerus fractures, rather than nonunions. Unfortunately, the current nonunion literature does not allow calculation of the rate of AVN or estimation of its clinical significance.
4.7.4 Osteosynthesis
4.7.4.1 ORIF with Standard Plates
Several studies reported good results when osteosynthesis was used for the treatment of surgical neck nonunions of the proximal humerus. Healy et al. retrospectively reviewed their experience and found better functional results after open reduction and internal fixation than after hemiarthroplasty , unreamed intramedullary implants, or non-operative management. Fixation techniques varied- the majority of their patients were treated with a 4.5 mm T-plate and a tension band through the rotator cuff, but others were stabilized using semitubular plates, dynamic compression plates, or Cobra plates. 12 of 14 (86%) nonunions achieved union at an average of 4.8 months postoperatively. Both nonunions that failed to unite were performed on the same patient who had significant medical comorbidities that may have contributed to her lack of healing. Autogenous bone grafting was performed in 12 of 14 cases. 11 of 12 grafted cases and 1 of 2 procedures performed without grafting achieved union. There were nine good results, four fair results, and one poor result. Shoulder range of motion averaged 110° of elevation, 33° of external rotation, and internal rotation to the thoracolumbar junction postoperatively [38].
4.7.4.2 ORIF with Blade Plates
Osteosynthesis using a blade plate has been shown to be successful for achieving union for nonunions of the surgical neck (Fig. 4.2a–i). Ring et al. reviewed 25 patients who underwent blade plating with autogenous iliac crest bone grafting, 23 of which (92%) united their fractures. Eighty percent of the patients obtained good or excellent functional results [39]. Allende and Allende presented their results of 7 patients with atrophic proximal humerus nonunions who underwent surgical treatment using a locking blade plate and all seven cases achieved union. It took an average of 5.9 months for patients to have radiographic evidence of union and patients who had allograft bone grafting took longer to achieve union compared with those who received autogenous bone graft (average of 7 vs. 5 months). Postoperative scores using the Disabilities of the Arm, Shoulder, and Hand (DASH) averaged 25 points and Constant scores averaged 72.7 [40]. Tauber et al. used a blade plate to treat 45 of 55 patients with proximal humerus nonunions who had sufficient bone stock in the humeral head. A blade plate was fashioned by bending a one third tubular plate and patients were not bone grafted. 41 of 45 patients (91%) achieved union. The four failures were attributed to varus collapse with screw pullout and underwent revision with a blade plate made from a 4.5 mm DC plate. Patients treated with blade plates in this study had an increase in Constant scores from 30.5 to 85.5 and had improvements in active shoulder flexion and abduction. Complications included two patients (4.4%) with surgical site infections and two patients (4.4%) who developed AVN [41]. Galatz and Iannotti treated 13 patients with nonunions of the proximal humerus: ten with blade plates and three with T-plates, combined with auto- or allograft. Eleven patients achieved excellent results and only one patient failed to achieve union. The persistent nonunion occurred in a wheelchair-dependent patient who returned to functional use of the involved upper extremity early and ultimately broke her T-plate but went on to successful union and excellent results following a revision surgery. Pain scores improved from 4.2 to 1.2 on a 5-point scale and the average shoulder flexion improved from 23.8° prior to surgery to 143.8°. All patients in this study achieved overhead elevation and were able to perform activities of daily living independently [13].
Fig. 4.2
A 37-year-old male sustained 3-part proximal humerus fracture (surgical neck and lesser tuberosity) during an motor vehicle accident. a Underwent open reduction and internal fixation with locked plating and allograft, seen on anteroposterior (AP) (b) and axillary (c) views. Went on to surgical neck nonunion with varies collapse and screw back out on AP (d) and axillary (e) views. Patient underwent hardware removal and revision fixation using a blade plate, fibular strut allograft (outlined), and bone morphogenic protein, seen immediately post-operatively in AP with internal rotation (f) and external rotation (g) views. Union was achieved at seven months on AP (h) and lateral (i) views
4.7.5 Use of Augmentation
4.7.5.1 Structural Graft Augmentation
The use of an intramedullary peg graft was first described by Walch et al. who treated 20 patients with proximal humerus nonunions using cortico-cancellous auto-graft harvested from the iliac crest, tibial crest, or fibula, in addition to osteosynthesis using a T-plate. Nineteen of 20 patients went on to union at an average of 4 months and demonstrated significant pain relief and improvement in shoulder flexion from 60° to 131°. The patients had six excellent and six good results, but 3 patients sustained tibial fractures following graft harvest and authors recommended against future use of this graft site [42].
The technique of using a fibular strut allograft as an intramedullary implant provides some inherent stability when the graft is impacted into the humeral head, and improves proximal screw purchase as described by Badman and Mighell, but without the graft donor site morbidity noted above. Badman and Mighell published the results of 18 patients who underwent intramedullary fibular strut allografting and stabilization of proximal humerus nonunions using locked plating. 17 patients (94%) achieved union at an average of 5.4 months. The remaining patient was a heavy smoker who had failed two previous attempts at surgical stabilization and ultimately required a hemiarthroplasty. Postoperative assessment of shoulder motion showed average flexion to 115°, external rotation to 37°, and internal rotation to the 10th thoracic vertebra. ASES scores improved from 40 preoperatively to 81 and analog pain scores improved from 6.7 to 1.5. Complications involved two posterior cord brachial plexus palsies that improved within 3 months and two cases of adhesive capsulitis, both requiring arthroscopic capsular release [26].
4.7.5.2 Biologic Augmentation
Augmentation with the recombinant human bone morphogenic protein rhBMP-2 can be used for nonunions when biological activity is felt to be lacking. The current literature supports the use of BMP for acute open tibia fractures [43], tibial nonunions [44], and recalcitrant long bone nonunions [45], but no studies have reported on its use for nonunions of the proximal humerus. A Cochrane review of BMP use for fracture healing in adults concluded that there is a paucity of data currently available and its role in treating nonunions remains unclear. Furthermore, they highlighted the high risk of bias in these studies due to industry involvement [46]. The U.S. Food and Drug Administration has granted rhBMP-7 (OP-1) a humanitarian device exemption (HDE) for treating recalcitrant long bone nonunions when autograft is unfeasible and alternative treatments have failed [47], but rhBMP-7 cannot be used “off label.” The FDA has approved rhBMP-2 for use in acute open tibia fractures, but its use for proximal humerus nonunions must be recognized as “off label.” Therefore, the use of biologic augments such as rhBMP-2 should be carefully weighed in light of the additional cost and unsubstantiated efficacy for treating proximal humerus nonunions compared to local or iliac crest autograft [8].
4.7.6 Unreamed Intramedullary Rods
The use of unreamed intramedullary implants (Rush rods) for the treatment of proximal humerus nonunions led to less favorable results. Five patients in a series presented by Healy et al. were treated using intramedullary implants: two with Rush rods alone, one using Rush rods with a tension band, one with an Ender nail, and one with a Lottes nail. Only one patient united their nonunion and all had poor functional results. The average shoulder motion was 40° of flexion and 10° of external rotation [38]. Duralde et al. presented a retrospective review of 20 patients with nonunions of the surgical neck, including 10 treated with open reduction and internal fixation and 10 with hemiarthroplasty based on an intraoperative assessment of the suitability of the fracture fragments for fixation. The fixation construct utilized Enders rods with a nonabsorbable suture or wire tension band and iliac crest bone grafting of the nonunion site with an onlay of cortical bone graft surrounding the nonunion site. Five of the 10 patients achieved union at an average of seven months. Of the remaining five, two were converted to a hemiarthroplasty, one underwent revision ORIF with a free fibular graft, one developed a deep infection and underwent removal of hardware and humeral head resection, and the final patient refused further surgical intervention. Two of the five that achieved union had an excellent result and three had a satisfactory result. Three later underwent reoperation for removal of painful prominent hardware. All five that failed to achieve union had unsatisfactory results [17].
Nayak et al. reviewed 17 patients who underwent operative intervention for nonunions of the surgical neck: 10 were treated with open reduction and internal fixation using Rush rods, tension banding, and bone graft, and seven had a hemiarthroplasty. Two of the 10 patients who had Rush rods had persistent lucencies on imaging indicating a failure to unite. The Rush rod group had a better average postoperative range of motion than the hemiarthroplasty group with more elevation (140° vs. 110°), but complications were common in this group. Intraoperative complications included one case of circumflex artery laceration, one permanent axillary nerve injury, and one cortical perforation with a Rush rod. Postoperatively, two patients had radiographic evidence of avascular necrosis of the humeral head and all had symptoms consistent with impingement of the Rush rods. Eight patients had hardware removal after achieving union [19]. No study has been able to reproduce the union rate of 92% obtained by Neer, who treated 13 patients using unlocked intramedullary implants with rotator cuff tension banding. All patients required a second surgery to remove their prominent, symptomatic hardware [21].
4.7.7 Interlocking Intramedullary Nails
The availability of more modern interlocked intramedullary nails has increased the union rates and decreased the rate of symptomatic hardware when intramedullary implants are used for the treatment of nonunions (Fig. 4.3a–h). Yamane et al. reviewed 13 patients who underwent surgical stabilization using an interlocking intramedullary nail with bone grafting. The intraoperative technique involved locking the nail with two proximal and two distal screws and emphasized seating the nail below the subchondral surface of the humeral head to avoid impingement. Patients with large bone voids underwent cancellous iliac crest autografting to fill the void, while smaller defects were filled with tricalcium phosphate cement. All patients achieved union without evidence of malunion or avascular necrosis. Japanese Orthopedic Association shoulder scores averaged 85 points postoperatively, with four excellent, seven good, and two fair results. Postoperative shoulder range of motion demonstrated flexion to 122° and external rotation to 35°. The only complication was the backing out of proximal interlocking screws, which required removal in 2 patients. It should be noted that 11 of 13 patients treated in this study had not previously undergone surgical treatment and the two who had been treated operatively underwent percutaneous pinning or intramedullary nailing, so it is unclear if this protocol would achieve similar results in patients who had previously undergone fixation with plating constructs [16].
Fig. 4.3
A 60-year-old female sustained a surgical neck fracture after a ground-level fall, seen on anteroposterior (AP) (a) and coronal computed tomography (b). Underwent open reduction and internal fixation with locked plating and allograft on AP (c) and lateral (d) views. Presented two years postoperatively with continued pain and immobility with varus collapse and hardware pullout on AP (e) and axillary (f) views. Patient underwent hardware removal and revision fixation, using an interlocked intramedullary nail, allograft, and bone morphogenic protein, with union at three months on AP (g) and axillary (h) views
4.7.8 Summary of Internal Fixation Devices
Technological advances in implant design offer the potential for improved union rates and decreased postoperative hardware prominence for nonunion patients treated with osteosynthesis. The need for adequate bone stock and a viable humeral head without significant glenohumeral arthritis has not been obviated by the availability of modern implants. Regardless of the implant used, the preparation of the nonunion site with resection of fibrous tissue and avascular bone is critical. Bone loss at the nonunion site and humeral head cavitation are commonly encountered challenges and autograft bone grafting is widely utilized to address these issues, although some authors have achieved union of proximal humerus nonunions using allograft, tricalcium phosphate cements, or without augmentation. Locking plate technology and intramedullary fibular strut allografts have lessened, but certainly not eliminated the difficulties associated with the osteopenia commonly encountered in surgical neck nonunions. A careful balance should be struck between the improved stability achieved through compression at the fracture site and the deltoid weakness associated with over shortening between the articular surface and the deltoid insertion. Varus alignment is often a progressiDeepika Uve deformity and should be carefully avoided. Newer interlocking intramedullary nails increase the stability at the nonunion site compared to Rush and Ender’s rods and may be inserted with decreased soft tissue stripping compared to plating constructs. These advances have served to increase the number of proximal humerus nonunions that can be treated with internal fixation rather than an arthroplasty.
Related posts:
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
