Background: The treatment of recurrent posterior instability of the shoulder, especially when it is associated with voluntary subluxation, remains controversial, and operative correction generally is not advised.
Methods: The results of operative correction of recurrent posterior subluxation in a consecutive series of twenty-six shoulders in twenty-four patients were reviewed. Eighteen shoulders were on the dominant side. The average age of the patients was twenty-four years (range, fifteen to thirty-three years). All of the patients had involuntary as well as voluntary posterior instability, but none had a psychiatric disorder. Only five patients had sustained a definite injury that had initiated the instability. Seven shoulders had had previous operations. A program of nonoperative treatment for a duration of at least three months had failed to control the symptoms in all patients. The twenty-six shoulders were treated with a posterior-inferior capsular shift procedure, which included repair of a so-called posterior Bankart lesion in seven shoulders. In addition, one of the shoulders had a posterior bone block and three shoulders (in two patients) had an osteotomy of the posterior part of the glenoid because of excessive glenoid retroversion. The outcome was assessed by means of a personal interview and a clinical examination, which included calculations of a score according to the system of Constant and Murley and the performance of the Simple Shoulder Test, and by means of a radiographic examination, with standardized radiographs and computerized tomography scanning.
Results: At an average of 7.6 years (range, 1.8 to 14.6 years) after the operation, the patients estimated that the function of the shoulder was an average of 86 percent of that of a normal shoulder. The average relative score according to the system of Constant and Murley was 91 percent. The subjective result was excellent for sixteen shoulders, good for eight, and fair for two. More than half of all of the patients were able to perform all activities of the Simple Shoulder Test, but eight patients (eight shoulders; 31 percent) still had discomfort at night. Five patients (21 percent) changed their profession because of the shoulder. All but one shoulder had a nearly normal active range of motion. The instability recurred in six (23 percent) of the twenty-six shoulders; three recurrences were in shoulders that had had a primary operation, and three were in shoulders that had had an operation on the posterior aspect of the shoulder before the index procedure. The instability did not recur in four shoulders that had had previous operations on the anterior aspect of the shoulder.
The subjective shoulder value, which was the patient's estimation of the value of the affected shoulder as a percentage of that of an entirely normal shoulder, was significantly higher for the stable shoulders (91 percent) than for the unstable shoulders (72 percent) (p < 0.05). The relative score according to the system of Constant and Murley was also higher for the stable shoulders (93 percent) than for the unstable shoulders (87 percent), but the difference was not found to be significant, with the numbers available. The joints were found to be well centered radiographically, and only six shoulders showed minimum signs of osteoarthritis. Computerized tomography scanning revealed an average glenoid retroversion of 3.2 degrees (range, 17 degrees of retroversion to 22 degrees of anteversion). When only the shoulders that had not had a posterior bone block or an osteotomy of the posterior aspect of the glenoid were considered, the average glenoid retroversion of those that had recurrent instability was 12.5 degrees, whereas it was only 6.2 degrees for those that remained stable (p < 0.05).
Conclusions: Overall, operative correction of voluntary posterior instability of the shoulder yielded very satisfactory intermediate-term clinical results. Recurrence was associated with a previous operation on the posterior aspect of the shoulder or with a new traumatic injury of an involved shoulder on the dominant side. The prevalence of recurrence did not increase over time, and clinically detectable osteoarthritis did not develop.