Two hundred and eighty-seven shoulders with a Bankart repair of the anterior shoulder instability were identified from January 2000 through to December 2002. The preoperative medical records and other relevant information from each patient were retrieved and reviewed. Those who had associated fractures of the proximal humerus, rotator cuff tear, previous shoulder operations, and glenohumeral or acromioclavicular arthritis were excluded. In total, two hundred and forty-six shoulders in two hundred and thirty-eight patients were examined. There were 214 men and 24 women. The average age at the time of surgery was 25 years (range, 14 to 47 years). Arthroscopic repair and open repair was performed in 130 and 116 shoulders, respectively.

A retrospective review of the prospectively collected data was conducted. The preoperative patient chart, radiographs, surgical records, and intraoperative arthroscopic photographs of all patients were analyzed. Age at the time of the first dislocation, the frequency of dislocation, interval between the first and second dislocation, activity of the patient and intraarticular pathologies observed during surgery were identified as prognostic factors that can affect the progress of traumatic dislocation. The relationships between factors were also analyzed. All procedures were performed by the same surgeon. The age at the time of the first dislocation, frequency of dislocation and interval between the first and second dislocation were analyzed by reviewing the surgical; medical records and by a telephone interview. Along with a reassessment of the medical records and an analysis of the arthroscopic photographs, the intraarticular pathologies were classified as classic Bankart lesions or non-classic Bankart lesions. Non-classic Bankart lesions included: Bony Bankart lesions, humeral avulsion of the glenohumeral ligament (HAGL),15) an anterior labroligamentous periosteal sleeve avulsion (ALPSA)-like lesions,16) and midsubstance tears. The manifestations of each lesion were examined to determine if it affected the natural history of shoulder dislocation. The presence of bony Bankart lesions, shown in preoperative radiological studies, and the size of the Hill-Sachs lesions were also considered to be contributing factors.

A dislocation of the shoulder is the most common of all joint dislocations and a recurrent dislocation is the main complication after an acute traumatic anterior dislocation of the shoulder. Hovelius17,18) reported a 1.7% incidence of traumatic dislocation among the general population in Sweden and 8% among Swedish hockey players in the highest league. According to Simonet and Cofield6) and Norlin,19) the incidence in a group of 20 to 30 years individuals ranged from 13 to 18 per 100,000 inhabitants. Based on these and more recent studies, the redislocation rate ranges from 4% to 96% depending on a variety of factors, such as age, athletic activity and associated lesions.

Among those factors, the patient's age at the time of the first dislocation was found to be the most important contributing factor in redislocation. The rate of redislocation increased with decreasing age of the patients at the time of the first dislocation. McLaughlin and MacLellan3) found that 95% of 181 primary dislocation patients aged 11 to 20 years experienced a redislocation. Rowe et al.4) also found, in a series of 488 patients who had both a primary and recurrent dislocation, an 83% recurrence rate in 107 shoulders of patients who were younger than 20 years. Meanwhile, Hovelius et al.2) reported a 47% rate of redislocation at the 2-year follow up of 102 primary traumatic dislocation in the under 22 years group. At the 5-year follow up, the rate increased to 64% and 67% at the 10-year follow up. In a group of 60 shoulders in patients aged from 23 to 29 years, the redislocation rate was 28%, 48% and 58% at the 2-, 5- and 10-year follow up, respectively. In a group of 95 shoulders in patients aged 30 to 40 years, the rate of redislocation was 13%, 20%, and 21% at the 2-, 5-, and 10-year follow up, respectively. Although previous reports indicated that the rate of redislocation increased with decreasing patient's age at the time of the first dislocation, unsolved questions still remained regarding the relationships between age and the frequency of dislocation, as well as between the frequency and interval of redislocation.

In this study, the results showed that as a patient was younger at the time of the first dislocation, the frequency of a redislocation increased. Indeed, the group under 20 years experienced a 3 times higher frequency of redislocation than the over 20 years group. The intervals differed from 59 weeks in the over 26 years group to 18 weeks in the 16 to 20 years group. This large difference provides strong evidence that a dislocation in people under 20 years will redislocate with a short interval. On the other hand, patients who had longer intervals between the first and the second dislocation showed a comparably less frequent rate of redislocation.

Athletic activity is another indicator that can influence a redislocation. Simonet and Cofield6) found an 80% of redislocation rate in athletes under 20 years and 30% in non-athletes in the same age group. Wheeler et al.7) recommended arthroscopic treatment for younger athletes after the initial dislocation due to the 92% rate of redislocation in young athletes who were treated conservatively. Arciero et al.5) reported details of 36 shoulders of athletes (mean age 20 years) who had experienced traumatic first-time anterior dislocation. Eighty percent (15 shoulders) of the conservatively treated group experienced a redislocation while only 14% (21 shoulders) of the arthroscopic treatment group experienced a redislocation. Therefore, they reported that surgical treatment should be suggested to patients under 20 years who experience a dislocation in order to prevent recurrence. te Slaa et al.20) noted that recurrence took place in 64% of patients younger than 20 years of age and in 6% of those older than 40 years, but recurrence was unrelated to the activity of the patient.

In this study, athletes experienced their first dislocation at a comparatively younger age than non-athletes, and had a significantly shorter interval between the first and second dislocation. However, once a redislocation occurred, the rate of recurrence in athletes and non-athletes was similar.

The type of lesions, classic or non-classic Bankart lesions, had no relationship with the patient's age at the time of the first dislocation. It might be believed that athletes would have more severe lesions in the capsulolabral complex as the redislocation rates increased. However, the results showed that the presence of classic and non-classic Bankart lesions were unrelated to the redislocation rates. Another unexpected outcome is in the group with non-classic Bankart lesion and bony Bankart lesion. Due to the excess exercise and repetition of certain postures, it was believed that athletes would experience more frequent redislocations, resulting in non-classic Bankart lesions, particularly bony Bankart lesions. However, there was no difference in these categories between athletes and non-athletes.

In summary, the age, activity, frequency and interval of dislocation have close relationships with the redislocation rates, but the presence of non-classic Bankart lesions or bony Bankart lesions are unrelated to the redislocation rates. It is possible that the type of Bankart lesion and the severity of bony Bankart lesions are determined at the first dislocation. Remarkably, patients with bony Bankart lesions demonstrated a high frequency of dislocation. Therefore, patients who present with bony Bankart lesions in radiography are likely to have recurrent shoulder instability after the dislocation.

In order to predict the natural history of traumatic anterior dislocation of the shoulder, more objective and various cases will need to be analyzed. However, this study is limited by its retrospective nature. For further study into the natural history of shoulder dislocation, it will be important to examine the progress of shoulder dislocations related to the level of force causing the injury, how immobilization relates to the total number of dislocations and the interval between the dislocations as well as how adequate rehabilitation affects the progress of dislocation.

In this study, the age at the time of first dislocation and athletic activity were related to an early redislocation and a high frequency of dislocation. Bony Bankart lesions were found more often in patients showing a higher frequency of dislocation. Early surgical treatment would be a good option for athletic patients who are under 20 years if they present with bony Bankart lesions after the first dislocation and have a short interval of recurrence despite the conservative treatment.

It is possible that various lesions of anterior shoulder instability are determined at the time of the first dislocation because the type of lesions were less affected by prognostic factors, such as age, frequency of dislocation and interval between dislocations.