Decision-making in the management of traumatic posterior shoulder dislocation and subsequent posterior shoulder instability (PSI) is a complex process with varying approaches. On one hand, researchers like Blacknall and Moroder have shown good results with conservative therapy [7, 19]. On the other hand, Cruz-Ferreira et al. established the superiority of arthroscopic therapy for PSI in their direct comparison of treatment options [22]. Considering this discrepancy, our retrospective study aimed to directly compare conservative and operative treatments and analyze our therapy algorithm, although our initial intention was not exclusively arthroscopic treatment.

Our case–control study demonstrates that nonoperative treatment is a recommended approach for traumatic posterior shoulder dislocation, particularly in cases without a direct surgical indication and low initial clinical complaints, as indicated by a low Western Ontario Shoulder Instability Index (WOSI) score. Overall, 25 out of the 28 (89%) patients benefited from the treatment and could be classified as having achieved therapeutic success. We considered a clinical therapeutic success to be present when an improvement of at least 10.4% in the WOSI score was recorded, consistent with the recommendation of Kirkley et al. [23]. The success rate in the group of patients treated conservatively did not significantly differ from that of the group undergoing surgical treatment (p = 0.980). Among the patients treated with arthroscopic procedures, the average WOSI score dropped from 133 points before treatment to 53 points after the operation, representing an 80-point improvement. Of the 17 patients treated arthroscopically, 16 achieved successful outcomes, with ten of them having previously attempted conservative therapy. Conversely, patients who received conservative treatment improved their WOSI scores from an initial average of 122 points to 57 points (Table 1).

These results align with our initial hypothesis that conservative therapy, when indicated, can yield positive outcomes comparable to those achieved through operative therapy. Therefore, immediate surgical intervention cannot be justified, given the associated risks relative to conservative treatment. While the improvement in the WOSI score was greater in the operative treatment group, it is important to note that these patients typically began therapy with more severe trauma and higher initial WOSI scores. Nevertheless, only two patients in our study did not show significant improvement. One patient, numbered 24, who was treated conservatively, only experienced an 8% improvement. The exact reasons for this treatment failure remain unknown. In contrast, patient 18, initially treated conservatively, saw a marginal increase in their WOSI score from 179 to 199 points, representing an 11% difference. Patient seven, with an initial score of 201 points, received arthroscopic treatment, which resulted in a lower score of 179 points, also reflecting an 11% difference and representing an unsatisfactory outcome.

An analysis of therapy failures indicates that patient 18 refused arthroscopic surgery and opted for conservative treatment. After a telephone consultation, she reported worsening symptoms and hand numbness while cycling. In the case of patient seven, multiple arthroscopic procedures were performed, which, according to studies by Jain et al. have been associated with worse outcomes in patients with nontraumatic PSI [18, 24]. An error rate of 11% in our study is consistent with the average reported in comparable studies, including the Norwegian Register for Shoulder Instability Surgery and the systematic review by Leivadiotou et al. which documented a mean recurrence rate of 5.51% among 387 patients with PSI [2, 17]. An analysis of WOSI scores was conducted using paired t tests on different patient subgroups to compare the scores of arthroscopically treated patients with those who received conservative treatment. This analysis revealed a significant difference with p < 0.05.

In our study, we confirm the currently recommended arthroscopic treatment for patients with bone damage, and our arthroscopic cohort demonstrates excellent results. To our knowledge, only one study has directly compared arthroscopic and conservative treatments. Cruz-Ferreira et al. concluded that operative treatment yielded better outcomes than nonoperative therapy, though this conclusion was primarily based on slightly improved scoring systems [22]. Our study design was not explicitly aimed at comparing therapy alternatives directly, as it is a two-arm case–control study for PSI, with treatment changes occurring in ten out of 28 patients within the first three months. Our primary focus was the initial decision-making process, rather than the intention to directly compare treatment options.

In Cruz-Ferreira’s study, the comparison of therapy options showed that in the nonoperative and operative groups, the Constant score was 78 versus 87, the Rowe score was 64 versus 88, and the Walch-Duplay score was 69 versus 82, which closely mirrors our results [22]. However, the risks associated with operative intervention were not fully considered by Cruz-Ferreira et al. when assessing treatment options. Our conclusion supports the idea that conservative therapy can offer similar success with fewer risks in patients without absolute operative indications and low initial subjective complaints, as reflected in the WOSI score.

Minor structural injuries following PSI, such as capsule ruptures or minor SLAP lesions, can often be effectively treated with physiotherapy, specifically focused on muscle strength regain. According to a biomechanical theoretical study by Hölscher et al. the shoulder joint is primarily guided by muscles, and the humeral head is centered by the shoulder reaction force produced predominantly by the rotator cuff muscles [8]. Training these shoulder muscles to redirect the force toward the centre can support the healing process in cases of posterior shoulder complex trauma involving labral and capsule injuries [8]. The results of Moroder et al. who achieved positive outcomes using EMS training to strengthen shoulder muscles in patients with functional posterior shoulder instability, reinforce the efficacy of this approach [10]. In line with our findings, the study by Blacknall et al. demonstrates that specialized physiotherapy rehabilitation is a valuable treatment option for atraumatic posterior shoulder instability, resulting in significant clinically important improvements in patient-reported outcomes [7].

In our study, the Western Ontario Shoulder Instability Index (WOSI) was employed as a reliable instrument to measure treatment success, which is independent of the examining physician. Furthermore, the “Core Outcome Measures in Effectiveness Trials” (COMET) suggests that future shoulder studies should encompass pain, physical function, the overall assessment of therapy success, and health-related quality of life as their main categories [25]. The WOSI covers three of these four main criteria: pain, physical functionality, and health-related quality of life. It also provides the most accurate recording of subjective symptoms and is known for its reliability, validity, and accuracy [25, 26]. The greatest advantage of the WOSI is its ability to measure shoulder instability in relation to an individual’s quality of life, which is why it was selected for this study.

Many studies have suggested a correlation between an increased glenoid retroversion and the risk of developing dynamic PSI. For instance, Galvin et al. found that patients with posterior glenohumeral instability (PGHI) exhibited an average glenoid retroversion of − 8.16° during MRI analysis, whereas the control group had a retroversion of − 2.9° [27]. Owens et al. took this concept further by postulating that for every degree of increased retroversion, the risk of PSI increased by 17% [28]. In contrast, Yoo et al. found no significant correlation between PSI and glenoid retroversion [29]. In our study cohort, we measured an average glenoid retroversion of − 4.3° using the Friedman method and an average retroversion of − 17.53° using the “glenoid vault” method. A deviation of 6.3° in the Friedman method falls within acceptable measurement tolerances, as confirmed by Friedman et al. who reported a range of − 12 to 14° in their control cohort [19]. This suggests that the examined patient cohort did not exhibit osteoarthritic changes in the glenoid, which, in other studies, could be linked to increased retroversion [27, 28, 30]. We concluded that the PSI in our cohort primarily originated from traumatic events. Given the relatively low prevalence of PSI, accounting for approximately 10% of all shoulder instabilities, the sample size is insufficient to achieve significant statistical significance. Additionally, the study design only allowed for treatment observation, and group assignments were determined retrospectively.

Ethically, it is debatable whether patients requiring immediate surgery and those who may not need surgery can be included in a comparative study. Our observational study leaves the question open regarding whether patients with a WOSI score below 125 might experience even greater improvement with surgery.

The results of this study are in line with the existing scientific knowledge, suggesting that the best initial approach is to initiate a three-month trial of nonoperative therapy when there are no absolute surgical indications [3, 31, 32]. If nonoperative therapy proves unsuccessful, then secondary surgery, particularly in cases with clear surgical indications, is recommended. For post-traumatic PSI without significant hard tissue defects, which still retains the ability to actively stabilize the shoulder, nonoperative therapy should be the primary choice, unless the WOSI score is higher than 130. In the case of shoulder injuries in performance athletes, primary reconstruction via arthroscopy is preferable [14].