A sixty-three-year-old woman with oxygen-dependent obstructive pulmonary disease presented with right-sided pleuritic chest pain. Her surgical history included multiple surgical procedures in the right shoulder due to instability and rotator cuff arthropathy. An arthroscopic rotator cuff repair in the right shoulder, done in 1989, had failed, and a shoulder arthrodesis was subsequently attempted. Shoulder implants were removed in 1992. Although the patient was asymptomatic, she was noted to have a fibrous nonunion of the right shoulder and had a revision arthrodesis in May 1996. This arthrodesis was unsuccessful, and there was subsequent failure of the implants. The patient was not a candidate for repeat revision arthrodesis because of inadequate bone stock and medical comorbidities.
A routine chest radiograph made in 2003 showed a broken plate with screws remaining in the right shoulder region as well as a large-fragment screw on the left side of the thorax, which was from a failed left glenoid reconstruction performed in 1995 at the time of a left total shoulder replacement (Fig. 1). In 2005, the patient sought orthopaedic treatment because of pain in the left shoulder. Radiographic findings were unchanged from those on prior chest radiographs. Computed tomography (CT) examinations of the chest demonstrated that the left-sided screw remained extrathoracic and had not migrated from its position over the previous ten years. Thus, no treatment was indicated for this screw.
Anteroposterior chest radiograph made in 2003 demonstrates the attempted right shoulder arthrodesis with plate-and-screw fixation. The screw visible in the left hemithorax migrated from the glenoid component of a failed left total shoulder arthroplasty. Since the left-sided screw was extrapleural and stable, it did not need to be removed.
The patient returned in 2006 because of right shoulder pain. Physical examination, radiographs (Figs. 2-A and 2-B), and a CT scan revealed that the patient had a nonunion of the right shoulder arthrodesis with no substantial migration of the implants. A revision fusion was not possible because of the degree of bone loss shown by the CT scan. Because of the multiple medical comorbidities and lack of substantial implant migration over a ten-year period, additional surgery was not recommended and the patient was instructed to return for follow-up as needed.
Figs. 2-A and 2-B Anteroposterior and lateral radiographs of the right shoulder made in 2006.
A routine chest radiograph, made to monitor the patient's pulmonary status in 2007, showed no further implant migration compared with that seen on prior chest radiographs (see Appendix). However, in April 2009, the patient complained of new, intermittent, sharp right-sided chest pain. A chest radiograph revealed intrathoracic implant migration (Fig. 3). No imaging studies had been performed from 2007 until the patient returned for evaluation in 2009. A CT scan performed at that time confirmed intrathoracic screw migration, with one of the large-fragment screws now adjacent to the subclavian artery, vein, and brachial plexus (Figs. 4-A and 4-B). A venogram showed the head of the screw abutting the subclavian vein (Fig. 5). On physical examination, there were no neurovascular deficits. The decision was made to remove the screws on an urgent basis, but not on an emergency basis because of the patient's multiple medical comorbidities.
Anteroposterior chest radiograph made in 2009 demonstrates the migration of large-fragment screws into the right hemithorax.
Figs. 4-A and 4-B Computed tomography axial scans made in 2009 demonstrate the migrated screws.
Preoperative venogram made in 2009 demonstrates abutment (arrow) of the screw head on the subclavian vein.
During surgery, the superior screw was palpable posterior to the clavicle. A 4-cm incision slightly posterior to the clavicle was carried deeply into the supraclavicular fat pad. Careful dissection permitted retraction of the subclavian vein and allowed visualization of the brachial plexus adjacent to the screw. One branch of the brachial plexus was adherent to the fibrous capsule surrounding the screw. This was carefully isolated with vessel loops and retracted. The screw was then removed without difficulty (see Appendix).
The right intrathoracic screw was accessed through a transverse incision created in the subaxillary space. After the screw position was confirmed with fluoroscopy, the head of the screw was palpated between the second and third ribs laterally. The screw was firmly fixed in a fibrous capsule and could not be easily extracted. A large-fragment screwdriver was used to extract the screw out of the chest. Pulmonary pressures on the ventilator were checked frequently and remained stable throughout the procedure. A postoperative chest radiograph revealed no pneumothorax and successful removal of the migrated screws (Fig. 6).
Anteroposterior chest radiograph made after removal of right-sided screws.
The patient was seen for follow-up at two and eight weeks after surgery, and no pulmonary or neurovascular complications were identified. The patient reported that the right-sided chest pain had resolved, and the function of the right shoulder had returned to baseline; however, she continued to have right shoulder pain. Since she had no pulmonary complications after the screw-removal surgery, the remaining implants were removed from the right shoulder and there was complete resolution of the shoulder pain.
Kirschner wires and Steinmann pins used for fixation of the sternoclavicular joint, acromioclavicular joint, clavicle, and proximal part of the humerus have migrated to seemingly random locations in the body, including the right ventricle of the heart, the lung, the spine, the abdomen, and the eye6,7. Migrating implants can lead to substantial morbidity, including pneumothorax, arrhythmia, and arterial fistula. Hemiplegia has been reported with implant migration to the spine2.
Why implants around the shoulder have a propensity to migrate is not well understood. In the majority of case reports, a long duration in situ has been described as an important factor contributing to migration of an implant. In our patient, the screws were in place for a total of thirteen years. Some have argued that repetitive body movements (respiration and upper-extremity motion) in the shoulder girdle contribute to the migration of implants. Lyons and Rockwood hypothesized that the possible mechanisms were "muscular activity, respiratory excursion, capillary action, electrolysis, regional resorption of bone, gravitational forces, and the great freedom of motion of the upper extremity2." As was seen in our case, pneumothorax from migrating implants is a rare event even after they cross the left hemithorax; the rarity of a pneumothorax from migrating implants is likely due to the thick adhesions and capsule that form around the migrating implant9.
There are several case reports of migrating implants after shoulder surgery1-5, but none have described the migration of a large-fragment screw after a failed shoulder fusion, to our knowledge. Mazet reported intrathoracic migration of smooth Kirschner wires after fixation of an acromioclavicular joint and shoulder fusion4. One patient was asymptomatic, and the other complained of pleuritic chest pain and a pneumothorax. Both patients required implant removal.
Sethi and Scott reported on migration of a Hagie pin after acromioclavicular joint fixation, with the pin ending up adjacent to the subclavian artery5. With pin removal, the artery was lacerated and required ligation. It is prudent to obtain studies, as we did, to determine the proximity of the implant to the local vascular structures.
Aalders et al. described the transthoracic and intraspinal migration of a 10-cm smooth Kirschner wire after fixation of an acromioclavicular joint in a nineteen-year-old man9. The Kirschner wire had migrated across the apices of both lungs and through the spinal canal without injuring the spinal cord. Removal required a thoracotomy and partial laminectomy.
To our knowledge, our case report is the first to describe the migration of intact large-fragment screws, not wires, after a failed shoulder arthrodesis and serves to emphasize that any type of implant in the shoulder girdle should be used with caution. Complications from migrating implants can cause substantial morbidity. This case report also illustrates the point that patients with broken implants, whether pins or screws, should be monitored closely. The surgeon should consider having a vascular and/or thoracic surgeon available for assistance when removing migrated implants from the thorax. Close follow-up with clinical and radiographic examinations can help to avoid morbidity.