**4. Classification**

In 1983 Samilson and Prieto [14] reported 74 patients with OA after multiple dislocations or surgical stabilization. They defined the term Dislocation Arthropathy and the Samilson and Prieto Classification (SPC) and suggested three stages based on anteroposterior radiographic images:


To increase the classification accuracy according to Samilson and Prieto, Buscayret et al. suggested subdividing the severe OA stage into two stages: one with humeral osteophytes above 8 mm and a last one with the loss of the joint gap (Figure 2) [15].

**Figure 2.** Image modified from Buscayret et al. [15]. (**a**): humeral osteophyte <3 mm. (**b**): humeral osteophyte >3 mm and <8 mm. (**c**): humeral osteophyte >8 mm. (**d**): obliteration of glenohumeral joint space.

The Samilson and Prieto classification is radiographic, but can it draw clinical conclusions? For example, Kircher et al. correlated pain, active and passive range of motion with OA graded according to Samilson and Prieto, finding the primary clinical feature, pain, as the main indication for surgery, not related to radiological parameters. In addition, the increasing size of the caudal humeral osteophyte was associated with a decreased functional status in all planes [16].

SPC is based only on 2D examination. Recently, Link et al. found no correlation between SPC and Walch classification for primary OA. Therefore, understanding glenoid morphology in the axial plane is mandatory in the final stage of OA for correct implant selection. However, no validated classification has been published to assess Dislocation Arthropathy in the axial plane [17].

#### **5. Causes and Prognostic Factors**

Since the first descriptions of OA after instability were carried out on patient groups that underwent surgery in most cases, it was presumed that OA results from the surgery itself [18].

#### *5.1. Development of OA after Non-Operative Management of Shoulder Instability*

Hovelius found that with a follow-up of 225 patients with first-time dislocations after 25 years, the spontaneous process after first-time dislocation accompanies OA development [19]. In a prospective study of patients with first-time dislocation after conservative therapy, 16.1% of 106 patients with a single dislocation without recurrence after ten years developed OA. With and without recurrence and operation, 11% of 208 shoulders had slight

and 9% moderate or severe OA. The interesting point is that the shoulders with only one recurrence had similar rates of OA to those with recurrent dislocations or operations [20].

Hovelius [19], and Singer [21], conclude that the primary dislocation introduces the development of the OA and that later recurrences are in this regard of minor importance. Ogawa et al., however, found that the number of the dislocations/subluxations was significantly different between shoulders with and without OA [9].

Buscayret et al. analyzed the pre and postoperative radiologic processes of 570 patients that had undergone shoulder stabilization [15]. They found five factors with statistically significant influence on the development of a preoperative OA without operation: The age at the first-time dislocation and at the time of the operation, in each case with higher risk at higher age; bony defects at the front lower glenoid or at the humeral head as well as a rotator cuff tear. Kraus et al. evaluated the results of conservative treatment of acute anteroinferior glenoid fractures [22]. Intra-articular step-off amounted to 6 mm (mean 2 mm); nevertheless, no significant increase in the OA rate could be found after a mean follow-up of 26.4 months. Marquiera et al. evaluated 14 patients with large Bankart fractures (>5 mm) and dislocation >2 mm that underwent conservative treatment [23]. After a mean follow-up of 5.6 years, every shoulder was stable. Only two patients showed mild and one patient moderate radiographic signs of arthrosis. Finally, Weisser et al. recently published excellent results of nonoperative treatment of anterior glenoid rim fractures after primary traumatic anterior shoulder dislocation. In the cohort of 30 patients with a >5 mm anterior glenoid rim fracture, functional outcome was reported as excellent with a low rate of recurrent instability (3%) and a low rate of new-onset OA (23%). To achieve these outcomes well centered post-reduction humeral head was mandatory. Anterior subluxation after reduction might develop in recurrent instability and OA, and should be considered a contraindication for nonoperative treatment [23,24].

The risk of developing a severe OA for individuals that suffered a shoulder dislocation is 10–20 fold increased [25]. The risk factors after conservative and surgical therapy are partially congruen<sup>t</sup> (age, extended time until operation, bony defects, alcohol, smoker, hyperlaxity, high BMI, and increased age at initial instability event) [15,26–28].

#### *5.2. OA after Surgical Stabilization*

The incidence of OA after anterior surgical stabilization is stated between 12–62% [2,4,29,30]. In a prospective study with 41 patients that had at least two anterior dislocations and underwent arthroscopic transglenoidal suture, 12% showed radiographic changes after a follow-up of 52 months. There was a significant correlation between these changes and a worse clinical outcome. Patients with Bankart and Hill–Sachs lesions or other bony alterations on the preoperative images presented with a significantly worse functional outcome [4].

In a retrospective study, 30 of 39 patients that underwent open Bankart reconstruction could be examined after a mean follow-up of 29 years. Five patients were treated with total shoulder arthroplasty, and seven presented with radiographic signs of OA. Overall, the rate of OA was 40%. The authors, therefore, concluded that even though satisfying long-term results could be attained, the development of OA could not be stopped by surgery [3].

In another retrospective study after open Bankart reconstruction 33 of 53 shoulders could be evaluated after 15 years; 87% presented with no or minor radiographic signs of OA, 14 patients with minimal, and one patient with severe signs of OA. A significant correlation between the radiographic degenerative alterations and limitation of external rotation in 90◦ abduction depending on the time of follow-up could be shown. An influence of the limited external rotation in developing arthrosis was discussed but could not be proved [2]. In 2010, Ogawa et al. reviewed 163 patients undergoing the open Bankart procedure, finding that the development and progression of OA cannot be prevented by surgical intervention [10]. Most postoperatively detected OA developed already before surgery; nevertheless, the progression of postoperative OA was prolonged. Recent studies with over 20 years of follow-up still report satisfying outcomes. Moroder et al., with a mean

22-year follow-up after open Bankart repair in 26 patients, reported good clinical outcomes with minimal loss in the range of motion [31]. However, OA was found in up to 50% of patients and was associated with loss of external rotation, raising the question of whether the loss of external rotation was caused by OA or by overtightening the anterior capsule.

After Latarjet stabilization, 56 of 95 patients could be evaluated after a mean follow-up of 14.3 years. Three factors could be identified to be relevant for the development of postoperative OA: rotator cuff lesions, intra- or postoperative complications, and positioning of the coracoid to lateral. This last one is the most important prognostic factor. It was discussed that even though the rate of OA is quite high, stage I OA seems not to influence the postoperative outcome even after ten years [29]. Mizuno et al. conducted a retrospective review of 68 open Latarjet patients with a mean follow-up of 20 years. Of the 60 shoulders without OA preoperatively, 12 developed OA at final follow-up [32]. Overall, postoperative OA was mild, finding stage 1 in 14.7%, stage 2 in 5.9%, and stage 3 in 8.8% of patients. On the other hand, Gordins et al. report 65% of OA in 31 patients after 33 years of follow-up open Latarjet [33]. However, the technique implemented was the one described by May, and all patients were operated on before the modified Latarjet technique by Patte et al. [34]. Coracoid dimensions and standing up "May coracoid transfer" might influence these OA outcomes.

Comparing the rate of OA after Bankart and Bristow–Latarjet procedures after ten years of follow up Hovelius found a higher incidence after the Bankart procedure (16 of 26) than after the Bristow–Latarjet procedure (9 of 30). A recent meta-analysis suggests that the Latarjet procedure has a lower OA degree than other treatments, including non-operative treatment [35].

After glenoid reconstruction of significant bony defects using a J-graft, most relevant studies showed that there was no significant correlation between the number of dislocations and the rate of OA and that a significant influence of the performed surgery could not be found [26]. A recent follow-up of the cohort published by Moroder et al. shows excellent results regarding stability and function after a mean follow-up of 18 years [36]. However, OA was present in 74% of the patients now. Therefore, the development of dislocation arthropathy may not be prevented by this procedure.

Of 34 patients that underwent Weber-osteotomy, only four (9%) had no OA, nine (26%) had been treated with total shoulder arthroplasty. The increased internal rotation and the degree of arthrosis were statistically significant [37].

The rate of OA after the Eden–Hybinette procedure is always mentioned to be one of the highest [38]. In a retrospective study including 74 shoulders with a mean follow-up of 29 years after the Eden-Hybinett procedure, a recurrence occurred in 15 cases (20%) and OA in 35 cases (47%).

The ages at the time of the primary dislocation, surgery, and follow-up were mentioned as risk factors. Shoulders with signs of OA showed significant limitations of external rotation, even though most of them were subjectively satisfied [39]. Comparing 2- and 5-years follow-ups, the degree of limitation in the external rotation was not correlated significantly to the rate of OA. The rate of arthrosis was higher after primary dislocation at a higher age (above 23) [40]. Buscaryet et al. showed in 570 patients after surgical stabilization that lower degrees of OA remain without progress more often than higher degrees; 19.9% of the patients who had no preoperative signs of OA developed postoperative OA [15]. The lengths of follow-up and the number of preoperative dislocations were found as risk factors. The time to surgery, the degree of instability (luxation or subluxation), the level of sportive activity, and especially the type of surgery were found to have no significant correlation with the development of OA. Therefore, there was no difference found between the Latarjet-procedure and soft-tissue techniques. When comparing the three groups with equal follow-up, no significant differences could be found concerning the rate of OA.

After arthroscopic stabilization, there was a lower rate of OA but a lower time of follow-up [15]. Other authors found similar rates of OA both in open and arthroscopic procedures [7]. Boileau et al. reported an increase in glenohumeral OA incidence from 4% preoperatively to 17% postoperatively after arthroscopic stabilization [41]. Meantime, a couple of literature reports deal with the long-term appearance of OA after arthroscopic Bankart repair. Castagna et al. found mild (29%) to moderate (10%) arthrosis after a minimum of 10 years after arthroscopic Bankart repair, but degenerative changes of the glenohumeral joint had no significant effect on the clinical outcomes [42]. Kavaja et al. examined the radiologic and clinical occurrence of glenohumeral OA 13 years after arthroscopic Bankart repair [43]. OA was diagnosed radiologically in 68 percent but rarely caused subjective symptoms. Franceschi et al. found OA in 21.8% of the patients with no preoperative degenerative changes eight years after arthroscopic Bankart repair [44]. Finally, the latest arthroscopic Bankart repair cohort published by Plath et al. reports 69% of OA over a hundred patients with a mean follow-up of 13 years [45].

These studies show that postoperative OA in different degrees occurs, both in open and arthroscopic procedures. As risk factors in developing postoperative OA, bony lesions (Bankart and Hill–Sachs lesions), lengths of follow-up, concomitant lesions of the rotator cuff, intra- and postoperative complications, positioning of the coracoid to lateral, higher age at primary dislocation or surgery, and a long time to surgery are mentioned. In addition, loose or proud metal pieces (screws, staples) (Figure 3) could cause a progressive OA quickly [46]. Yeh and Kharrazi report a rare but dramatic complication following shoulder arthroscopy: post arthroscopic glenohumeral chondrolysis [47]. The articular cartilage undergoes rapid degenerative changes shortly after arthroscopic surgery. Although the etiology of post arthroscopic glenohumeral chondrolysis is not ye<sup>t</sup> fully understood, the pathophysiology is likely multifactorial.

After arthroscopic stabilization using screws positioned at the glenoidal rim, Tauber et al. found only one case in 10 cases of material impingement that made the removal necessary after two years [48]. Experimental studies show that a loss of the anteroinferior labrum reduces the contact area by 7–15% and increases the contact pressure by 8–20%, concerning the anteroinferior part of the glenoid even at 53%. A bony loss of 30% of the glenoid diameter increases the contact pressure at the anteroinferior part even at 300–400% [49]. Whether such a loss or step-off formation is relevant for instability and development of arthrosis seems to be dependent on a centered or decentered humeral head.

#### *5.3. Capsulorrhaphy Arthropathy as an Own Entity?*

Matsen et al. defined the term capsulorrhaphy arthropathy for patients who develop OA due to too strongly strained anterior capsules [50]. The strong harnessing of the anterior soft tissues, e.g., a Putti–Platt, or a too strongly strained Bankart operation, leads to compression and intensified shearing stresses on the joint surface that increase if the patient goes into external rotation. It is postulated that this mechanism develops in all operations where the external rotation is excessively limited [51].

Biomechanical and anatomical studies today offer evidence that a non-anatomical strain of the anterior capsule leads to an increase in posterior joint pressure, posteroinferior subluxation of the humeral head, and thus pain and the development of arthrosis [52]. In a cadaveric comparison of a front capsule strain with an anteroinferior capsule shift, it could be shown that during the strain of the front capsule the stability decreases and the external rotation and elevation are limited. That larger shear joint forces are necessary to reach the maximum elevation. In contrast, the anteroinferior capsule shift improves stability without limiting the external rotation or elevation [53] (Figure 4).

Gerber and Werner experimentally showed the effect of selective capsulorrhaphy on the translation and the passive range of motion [54,55].

On the one hand, these studies document that capsulorrhaphy arthropathy is biomechanically justified and permits, on the other hand, developing more anatomical stabilization operations. In retrospective studies, a decreased external rotation was connected with an increased rate of OA; whether this was the cause or the effect could not be clarified [15].

**Figure 3.** Details of proud screws after glenoid fracture surgery removed arthroscopically. **First row**: radiologic studies showing instability surgery performed and proud implants. **Second row**: Images of a humeral cartilage defect and debridement necessary to expose implants. **Third row**: Arthroscopic screws removal.

**Figure 4.** 43 year-old patient treated with hemiprosthesis after capsulorrhaphy arthropathy subsequent to open instability repair. **First row**: preoperative radiographs. **Second row**: intraoperative pictures of hemiprosthesis implant and postoperative radiographs. **Third row**: Physical examination and shoulder function at final follow-up.

#### **6. Treatment Options and Results**
