Glenohumeral arthrodesis utilizing intraoperative computer navigation: a case report and surgical technique

Glenohumeral arthrodesis is a viable surgical option for patients with intractable pain and loss of function of the shoulder who are poor candidates for joint replacement or other salvage procedures.^2,3,6,9,18 The most common indications for glenohumeral arthrodesis are flail shoulder due to brachial plexus injury, with other indications including axillary nerve palsy, secondary osteoarthritis, persistent shoulder instability, and infection.^5,17 In cases of obstetrical brachial plexus injury, muscular imbalance about the shoulder may result in persistent posterior dislocation or subluxation of the humeral head and resultant glenoid bone loss, which can limit surgical treatment options.^3,15 For these patients, intraoperative computer navigation can assist with precise placement of screws and potentially reduce complications related to screw placement.¹¹ In this case report, we review a surgical technique for glenohumeral arthrodesis for a case of chronic posterior dislocation arthropathy and severe glenoid bone loss utilizing preoperative computed tomography (CT)-based computer navigation. This technique allows for precise screw trajectory and optimized arthrodesis construct in the setting of severe deformity.

Case presentation

A 68-year-old right-hand-dominant female with a history of obstetrical brachial plexus injury presented to our clinic with many years of severe right shoulder pain and stiffness. She had previously undergone proximal humerus derotational osteotomy at age 16, followed by subsequent hardware removal. On exam, she had significant limitations to her shoulder range of motion, which was primarily scapulothoracic and consisted of 60 degrees forward elevation, 10 degrees external rotation at side, and internal rotation to back pocket. She had 4 out of 5 isometric strength of her rotator cuff muscles. She had axillary nerve allodynia preoperatively with intact motor and sensory in radial, median, and ulnar distributions. Additionally, she had a chronic elbow contracture of 30°, which had developed secondary to her obstetric brachial plexus, but was able to achieve full elbow flexion. Imaging, including X-ray and CT scan, demonstrated severe secondary glenohumeral arthritis with significant glenoid retroversion, posterior glenoid bone loss, and chronic posterior dislocation of the humeral head with 100% subluxation in relation to Friedman's line. Because the patient had lived with a posterior dislocation since birth, the glenoid did not develop normally and defies classification by the standard systems such as Walch or Favard but would be closest to a Walch type C glenoid with approximately 36 degrees of retroversion using Friedman's line as a reference (Fig. 1). She was initially managed conservatively with physical therapy, pain medication, and a fluoroscopy-guided corticosteroid injection, which provided only 2 weeks of pain relief. After failing conservative therapy, we discussed surgical options, including hemiarthroplasty, reverse total shoulder arthroplasty, resection arthroplasty, and glenohumeral arthrodesis. Throughout her life, the patient had never been able to move her shoulder through a functional range of motion due to the brachial plexus birth palsy and resulting posterior shoulder dislocation. She was not particularly concerned about improvement in range of motion; her primary goal was elimination of the pain that the secondary arthritis was causing. We discussed surgical options including reverse total shoulder arthroplasty, hemiarthroplasty, arthrodesis, and resection arthroplasty as well as the risks and benefits of each option. Specifically, we discussed the increased risk of complications with shoulder arthroplasty secondary to her obstetric brachial palsy, lifelong posterior dislocation, and resulting severe glenoid deformity. These include failure of glenoid fixation, postoperative instability, scapula fracture, and inability to improve functional range of motion.^10,16,20 We also discussed that while there are several reports of cases of obstetric birth palsy successfully treated with shoulder arthroplasty in terms of pain relief, range of motion changes following arthroplasty have been modest, with some studies reporting no change and others reporting modest changes in forward flexion and external rotation but decreases in internal rotation following surgery.^10,16,20 Risks unique to fusion discussed included pseudoarthrosis, fracture, and hardware failure. She opted to proceed with arthrodesis after a thorough discussion of the risks and benefits of all the options.

Figure 1.

Preoperative XR and axial CT demonstrating severe glenohumeral arthritis with glenoid retroversion and posterior subluxation of the humeral head. CT, computed tomography; XR, X-ray.

Due to her significant glenoid deformity, the decision was made preoperatively to use intraoperative computer navigation to assist in optimizing screw trajectory, minimize risk of neurovascular injury, and maximize bony purchase (Fig. 2).^1,7,17 There was limited glenoid bone stock, and screws orthogonal to the neoglenoid face would need to be directed anteriorly, potentially placing the neurovascular structures of the brachial plexus and axillary artery at risk if drill or screws were too long or malpositioned. Prior studies have described the axillary artery as approximately 1-1.8 cm from the inferior glenohumeral margin, while the brachial plexus is within 2 cm and has branches as close as 5 mm to the margin in some anatomic studies.^13,19

Figure 2.

Preoperative plan for screw trajectory based on patient's CT scan. CT, computed tomography.

Surgery was performed in the beach chair position. A curvilinear incision was created from the posterior scapular spine across the lateral aspect of the acromion and down the anterolateral aspect of the arm, incorporating her prior healed anterior incision. A standard deltopectoral approach was utilized, and the anterior and middle deltoid were additionally released from the acromion to increase exposure posteriorly. Intraoperative motion was assessed, and only micromotion at the glenohumeral joint was observed. All shoulder motion was found to be scapulothoracic, but there was enough micromotion through the arthritic articulation that this was felt to be the primary pain generator.

At this point, we proceeded with intraoperative navigation for which we used the Medacta NextAR system (Medacta, Castel San Pietro, Switzerland). Two small pins were placed in the coracoid to which we affixed the tracker for the navigation system. The patient's scapula was registered with the preoperative CT scan using points on the coracoid, acromion, and glenoid to facilitate screw placement for the fusion (Fig. 3). The humeral head, greater tuberosity, undersurface of the acromion, and the glenoid were decorticated with a bur. The arm was placed in the chosen position of approximately 15-20 degrees flexion, 15-20 degrees abduction, and 25-30 degrees internal rotation, similar to her preoperative position since this was an in situ fusion and she had adapted functionally to the preoperative position of her arm. The joint was provisionally pinned in situ in accordance with our preoperative plan to maximize bony contact before placing the hardware. Utilizing computer navigation, three cannulated screw guide pins were placed from posterior-lateral to anterior-medial and measured off the navigation screen and confirmed with a depth gauge. Partially threaded cannulated screws (6.5 mm diameter) supplemented with washers were utilized given the poor bone quality. Bone graft was taken from the posterior aspect of the dislocated humeral head and packed into the arthrodesis site. Two additional screws were placed through the acromion into the humeral head under fluoroscopic guidance (Fig. 4). Final fluoroscopy confirmed appropriate position of all screws. The deltoid was reattached through drill holes in the acromion to preserve contour and soft tissue coverage.

Figure 3.

The tracking system consists of a tracker which is affixed to the coracoid and communicates via infrared sensors to a camera on the instruments. The control unit receives information from the tracking system wirelessly and integrates this information with the preoperative CT scan. Courtesy of Medacta.¹⁴CT, computed tomography; NextAR, augmented reality; TS, targeting system.

Figure 4.

Left—Intraoperative photo demonstrating approach and placement of screws using navigation system. Right—Image demonstrating control unit rendering of screw trajectory based on information from the tracking system. AP, anterior posterior; REB, relativistic electronic beam.

Postoperatively, the patient was immobilized in a sling for 6 weeks. She was seen at 2 weeks, 6 weeks, 3 months, 6 months, and 1 year postoperatively. She had near-complete pain relief by her 6-week visit; range of motion was unchanged from the preoperative range. Her preoperative Visual Analog Scale (VAS) score was 6, and at 1-year follow-up, she reported a VAS score of 0. Serial imaging demonstrated excellent consolidation of her glenohumeral joint fusion (Fig. 5). At her 1-year visit, she was back to all of her preoperative activities, such as kayaking and paddleboarding, without complaints related to her shoulder.

Figure 5.

Immediate postoperative XR (top row) and one-year follow-up XR (bottom row) demonstrating screw placement and solid fusion at follow-up. XR, X-ray.

Discussion

Arthrodesis has long been a treatment option for patients with secondary arthritis due to plexus palsy. Chammas et al compared functional outcomes after glenohumeral arthrodesis for patients with partial (11 patients) vs. complete (16 patients) brachial plexus injuries, reporting the major determinants of outcome following fusion to be hand function, position of fusion especially internal rotation, and pain, though a nonfunctioning hand is not a limiting factor for recovery of satisfactory active range of motion in this series.⁴ Lenoir et al reported a case series of 8 patients with brachial plexus injuries who underwent arthroscopic arthrodesis of the shoulder following motorcycle accidents, demonstrating glenohumeral fusion in all cases with a mean time to fusion of 3 months.¹² In this series, surgery was performed in the beach chair position with an external fixator that remained on for 2 months postoperatively.

In this case report, we demonstrated the use of intraoperative computer navigation for successful glenohumeral arthrodesis in a patient with obstetrical brachial plexopathy and painful secondary arthritis due to chronic posterior dislocation. While several case reports and series have detailed the successes and complications of glenohumeral arthrodesis, only one other case report exists to our knowledge where intraoperative navigation was used to assist in fusion construct. In that case, Lädermann and Denard used O-arm navigation with the patient in the lateral decubitus position to achieve glenohumeral arthrodesis. Their patient was a 52-year-old female with similar pathology to our patient, an obstetrical brachial plexopathy, severe retroversion, and posterior glenoid wear.¹¹ Their technique requires lateral positioning to allow for both O-arm use intraoperatively, whereas in our technique, intraoperative CT is not required since preoperative CT is used to create a plan which is then executed intraoperatively after performing radiation-free scapular registration. Moreover, this technique is possible in the beach chair position. Rojas et al reported the accuracy of this navigation system to be within 0.9° ± 0.8° for inclination and 1.2° ± 1.1° for retroversion between the intraoperative position reported by navigation software and postoperative position measured by CT scan.¹²

Intraoperative navigation has been used successfully in shoulder surgery most notably in shoulder arthroplasty to assist in positioning of implants, and is especially useful in cases with severe glenoid deformities.^7,8 Adapting this technique to glenohumeral arthrodesis allows for more precision in the placement of screws and may prevent complications associated with aberrant screw placement, especially in cases of severe glenoid deformity.

There are limitations to this study. While intraoperative CT-guided computer navigation holds promise for improving accuracy in surgery, this technology is not readily available at many institutions and may add costs to the procedure.¹⁷ Also, while this study establishes the feasibility of use of computer navigation for assistance in glenohumeral arthrodesis, it is a single case report. Additional experience is necessary before this would be considered a recommended technique.

Conclusion

Intraoperative computer navigation is a reproducible technique for glenohumeral arthrodesis in cases of severe glenoid bone loss and deformity, with the potential for more precise screw placement, improved strength of fixation, and reduced risk of hardware-related complications.

Disclaimers:

Funding: No funding was disclosed by the authors.

Conflicts of interest: John-Erik Bell received royalties and consultant payments from Medacta Inc., which is related to the subject of this work. The other authors, their immediate families, and any research foundations with which they are affiliated have not received any financial payments or other benefits from any commercial entity related to the subject of this article.

Patient consent: The patient provided informed consent for treatment and use of case materials for research purposes.