A fifty-five-year-old male office worker had undergone a modified Bristow-Latarjet procedure twenty years earlier because of anterior dislocations of the left shoulder. He had recovered fully from the procedure and had not experienced any complications.
While using his arms to climb out of a swimming pool, the patient felt something “pop” in the left arm (the same side as the prior Bristow-Latarjet procedure). He developed symptoms of shoulder pain and weakness in the arm and hand, which was associated with reduced sensation. On examination, he had reduced power of the deltoid and triceps muscles as well as weakness of wrist extension. Sensation to light touch and pinprick examination was reduced over the posterolateral parts of the forearm and dorsum of the hand in the distribution of the radial nerve. The examination was otherwise unremarkable. These symptoms and examination findings were consistent with injury of the posterior cord of the brachial plexus.
Radiographs (Fig. 1) and subsequent computed tomography (CT) (Figs. 2-A, 2-B, and 2-C) showed that the screw holding the original coracoid graft had become detached from the glenoid; it had migrated medially and inferiorly, presumably into the brachial plexus.
The patient was taken to the operating room to have the screw removed. This procedure was performed through an extended deltopectoral approach. Dissection was kept to a minimum, with no exploration medial to the conjoint tendon or release of the pectoralis major tendon. The screw was carefully removed from its close proximity to the brachial plexus and axillary artery, and there was no obvious adjacent scar tissue. On inspection, the screw had a small fragment of bone embedded in its threads (Fig. 3).
On immediate postoperative examination, it was noted that the muscle power had partially improved, but the sensory disturbance remained unresolved. The patient was discharged home the next day. At the follow-up six weeks after a course of physical therapy, there was full resolution of both the weakness and reduced sensation. One year postoperatively, he had not had any additional complications.
A recognized complication of surgery about the glenohumeral joint9 is screw migration, which can lead to various consequences. Three previous case reports have described postoperative neurological complications that presented between four months and three years after the initial procedure5-7. In all three cases, the screw was subsequently removed. In two of these cases, false aneurysms that had been created by damage from the migrating screw were repaired. In one case, a prominent part of the glenoid was trimmed, and the musculocutaneous nerve was dissected free from scar tissue. Symptoms fully resolved in only one of the three cases; the other two patients had improvement, but still had long-term neurological deficits.
Our patient provided a history of an abnormal “pop” while applying a forward traction force to the shoulder girdle while climbing out of a swimming pool. Intraoperatively, the screw, which had a piece of cancellous bone attached, was removed intact. This suggested that the screw had already partially migrated out of the glenoid, and this force had caused loosening of the remaining threads.
Previous studies have shown that screw loosening may contribute to almost 50% of all failed procedures about the glenoid involving screws9,10. It seems likely that this was the natural history of the complication in our patient. Coracoid nonunion is a relatively common complication and may increase the risk for screw displacement10.
When patients who have had a Bristow-Latarjet procedure present with neurological symptoms, a migrating screw should be a differential diagnosis, even many years after the initial procedure (as in our patient). It is important to obtain adequate radiographs when evaluating patients with shoulder injuries. It is very likely that a patient who had had surgery twenty years earlier might not even know or remember that there is an implant at the operative site. In addition, the initial evaluating physician may be unfamiliar with shoulder surgery and therefore may neglect to obtain a radiograph. A migrating screw should be removed to prevent additional damage.
We recommend that patients who have had a Bristow-Latarjet procedure should be followed with postoperative radiographic evaluation in order to prevent complications that can be caused by a migrating screw.