Does glenoid version affect the post-operative clinical outcome after repair of posterior labral tear: A retrospective study

Abstract

Introduction

Posterior shoulder instability and consequent labral tear have been predominantly associated with retroversion of the bony glenoid and chondrolabral version. However, literature concerning the degree of glenoid and chondrolabral version and clinical outcomes following labral repair is scarce.

Methods

A retrospective cohort study was undertaken among patients with posterior shoulder instability who had undergone arthroscopic isolated posterior labral repair. The MRI was used to assess the bony and labral variations of the glenoid using the Friedman method and further categorized as either anteverted or retroverted glenoid. American Shoulder and Elbow Surgeons (ASES), Oxford Shoulder Score (OSS), and Single Assessment Numeric Evaluation (SANE) scores were used to evaluate pre- and postoperative clinical outcomes at a minimum follow-up of one year.

Results

33 shoulder MRIs performed at our institution were available for radiographic analysis. Among the 33 shoulders that underwent posterior shoulder capsulolabral repair, 23 were available for clinical follow-up for at least one year (range, 12–108 months). The mean (±SD) labral version and bony version were −3.98 (±5.68) and −2.83 (±5.20), respectively. The mean (±SD) postoperative ASES, Oxford score, and SANE scores for all participants were 84.04 (±14.14), 43.38 (±3.81), and 78.57 (±17.40), respectively. However, the difference in clinical outcomes in patients with anteverted or retroverted glenoid versions was not statistically significant.

Conclusion

Our study concludes a strong correlation between posterior glenohumeral instability and higher degree of glenoid retroversion. Nevertheless, variations in the glenoid version appear to have no significant impact on clinical outcomes.

Level of study

Level IV, retrospective cohort.

1. Introduction

Compared to anterior instability, posterior shoulder instability, and associated posterior labral tears are relatively uncommon. Its incidence varies between 2 % and 12 % of all documented cases of shoulder instability.^1,2 Unlike anterior instability, wherein subluxations and dislocations are the predominant presentation, posterior shoulder instability can manifest in various ways, such as vague pain at the back of the shoulder along with decline in shoulder's strength or endurance, an inability to throw or decrease the velocity of throwing, or rarely frank subluxation-dislocation.3, 4, 5 Posterior shoulder instability arises due to repeated actions, such as throwing, push-ups, or bench presses, leading to microtrauma of the posterior capsulolabral complex, causing labral tear and capsular attenuation.⁶ This pattern of recurring microtrauma puts the individual at risk for posterior labral tears and capsular attenuation.⁷ Besides the aforementioned etiological variables, several studies have reported a correlation between posterior shoulder instability and retroversion of the bony glenoid and chondrolabral version.7, 8, 9 Although the correlation between the posterior shoulder instability and the glenoid version has been extensively studied, literature concerning variability in the glenoid version and its effect on clinical outcomes is still scarce.^10,11

The objective of this study is to ascertain the glenoid version (both bony and labral) in patients experiencing symptomatic unidirectional posterior instability and to evaluate the relationship between the glenoid version and the clinical outcome following arthroscopic posterior labral surgery.

2. Materials and methods

2.1. Patient selection

After ethical committee approval (IEC/396/2022), a retrospective cohort study was performed on patients with posterior shoulder instability who underwent arthroscopic posterior labral repair from January 2014 to December 2021. The study enrolled participants who met the following criteria: 1) aged between 18 and 50 years, 2) diagnosed with an isolated posterior labral tear using Magnetic resonance imaging (MRI), 3) had arthroscopic posterior labral repair, and 4) a minimum of one year of clinical follow-up. Their baseline demographic and clinical data were accessed from medical records. The study excluded individuals with anterior or multidirectional instability who underwent concomitant rotator cuff repair, a posterior bone lesion, or glenohumeral arthritis.

2.2. Patient evaluation

The American Shoulder and Elbow Surgeons (ASES) scores, Oxford Shoulder Score (OSS), and Single Assessment Numeric Evaluation (SANE) scores were utilized to grade the clinical outcomes pre- and post-surgery. ASES, Oxford, and SANE scores were obtained based on the latest follow-up from the clinical case records from the institute's Medical Records Department (MRD).12, 13, 14

2.3. Operative treatment

All patients were operated in a floppy lateral position under general anaesthesia with a scalene block. A single senior shoulder surgeon operated on all patients. A standard posterior portal was established, and diagnostic arthroscopy was performed. An anteroinferior portal was created just above the superior margin of the subscapularis tendon, and systematic probing of the joint was performed. Then, the arthroscope was shifted to the anteroinferior portal to confirm the extent of the posterior labral lesion. The labral lesions were stabilized with two to four PEEK or all-suture anchors anchors in standard fashion. No or minimal posterior capsule incorporation was done while repairing the labrum to avoid postoperative tightness in the posterior capsule, resulting in diminished internal rotation. Notably, several cases involved fraying of the articular surface of the posterior supraspinatus tendon, which was debrided using a radiofrequency device (Vapr, Mitek, Johnson and Jhonson, USA) and power shaver (Stryker, USA).

2.4. Postoperative rehabilitation

Post-operatively, the shoulder was supported in an arm sling, which was continued for four weeks. During the first four weeks, patients were allowed forward flexion up to 90°, abduction up to 90°, external rotation up to 20°, and isometric strengthening of rotator cuff and scapular muscles. After fourth postoperative week, a systematic increase in the range of movement (ROM) was implemented, aiming to achieve full ROM by the end of the 12th −16th postoperative week. Strengthening of the rotator cuff was initiated after six weeks. Sports-specific rehabilitation was initiated at the end of the 16th week, and return to sports was permitted once the patient regained more than eighty percent of the normal side's endurance and strength.

2.5. Radiographic evaluation of bony and labral versions of glenoid

The Magnetic Resonance Imaging (MRI) of all patients included in the study underwent MRI at our institution either on a 1.5 T Philips (Amsterdam, The Netherlands) or GE HealthCare scanner (GE HealthCare, Chicago, USA). The radiographic evaluation of the version involved the assessment of the glenoid bony version with the modified Freidman method and the chondrolabral version using Kim's method.^8,15

The modified Freidman method was used to measure the glenoid version on the axial section at the inferior third of the glenoid rim.¹⁵ The Friedman line, extending from the tip of the medial border of the scapula to the center of the glenoid fossa, defined the scapular axis. A line perpendicular (90°) to the Friedman line was established as the neutral glenoid version line. Another line was drawn between the anterior and posterior margins of the subchondral glenoid (Fig. 1). The measured angle of the bony version was the angle formed between the line of the neutral version and the line connecting the anteroposterior edge of the subchondral glenoid. The glenoid is considered retroverted if the posterior edge of the glenoid is positioned medial to the neutral version line. In contrast, the glenoid is anteverted if the anterior glenoid edge is medial to the neutral version line. Alternatively, positive and negative angles defined anteversion and retroversion, retrospectively.

Fig. 1.

Measurement of bony version. A-B is the Friedman line. C-D is a line perpendicular to the Friedman line, which represents the line of the neutral glenoid version. E-F is the line joining the anterior and posterior tip of the subchondral glenoid bone. “I” marks the intersection of all the lines. The angle formed by CIF or EID is the measured bony version.

The labral version, as measured by Kim's method, is defined by the angle formed between the line of the neutral glenoid version and a line connecting the apex of the anterior and posterior labrums (Fig. 2).⁸

Fig. 2.

Measurement of labral version. A-B is the Friedman line. C-D represents the line of the neutral glenoid version. E-F is the line joining the anterior and posterior tips of the labrum. “I” marks the intersection of all the lines. The angle formed by CIE or DIF is the measured labral version.

If the version measured neutral (0°), it was combined with the anteversion group for statistical analysis.

Utilizing the PACS (Picture Archiving and Communications System, Meddiff Technologies, 4.1.0), every measurement was made using a mouse pointer and an automatic computer angle calculator. Unaware of the clinical outcomes, two surgeons performed the measurements twice for all patients on different occasions to assess intra- and interobserver observer reliability. A third surgeon who remained unaware of radiological findings assessed the clinical outcome at the final follow-up.

2.6. Statistical analysis

SPSS v23 (IBM Corp., USA) was used for data analysis. The continuous variables were summarized using descriptive statistics such as means, standard deviations, and medians. The Wilcoxon-Mann-Whitney U Test was employed to compare the results of the anteverted and retroverted glenoid. Fisher's exact test was used to assess the association between the ‘Labral Version’ and the ‘Bony Version.’ Interclass Correlation Coefficient (ICC) was used to assess the intra- and inter-observer reliability of measuring the glenoid version.

3. Results

3.1. Baseline data

38 shoulders were operated on between 2014 and 2020; 33 received an MRI of the shoulder at our institution and were available for radiological analysis. Of the 33 shoulders with MRI, 23 were available for clinical evaluation, as ten patients were lost on follow-up and did not report back after six months of rehabilitation. One of the patients had bilateral infliction, but not at the exact moment. The baseline demographic and clinical data are mentioned in Table 1. Most patients were male (81.8 %), with a mean age of 29.16 years (Range, 19–51). 69.6 % of patients had atraumatic onset of shoulder pain. All 23 shoulders presented with posterior shoulder pain, while 10 patients (43.47 %) reported associated clicks. Only one had repeated subluxations. All 23 patients reported a loss of power during throwing activities.

Table 1.

Baseline data of all 23 shoulders (22 patients) who were clinically assessed. RTA- Road traffic accident. As one of the patients had bilateral infliction, the total number of individuals summed to 22, while the whole shoulder investigated was 23. RTA, road traffic accident; SD, standard deviation.

Sample characteristics	n	%	Mean	Range	SD
Gender
Male	18	81.8
Female	4	18.2
Age			29.14	19–51	9.39
Side affected
Right	16	69.5
Left	6	26.1
Bilateral	1	4.34
Traumatic (total)	7	30.4
• RTA	2	8.7
• Forceful jerk to shoulder due to a fall	4	17.4
• Fall of heavy object on shoulder	1	4.3
Atraumatic (total)	16	69.6
Sports-related	13
•Cricket	6	26.2
•Badminton	4	17.4
•Gymnasium injury	3	13.0
Insidious onset	3	13.0
Duration of symptoms before surgery (months)			13.95	3–60	15.72
Follow-up (months)			50.25	12–108	26.64

3.2. Radiological version

The glenoid version (labral and bony) was calculated for all 33 shoulders that had done MRI studies at our institution (Table 2). The prevalence of retroverted glenoid was higher in both the labral and bony versions (66.7 % and 57.6 %) [Table 2]. The mean (±SD) labral and bony versions of all 33 shoulders were −3.98 (±5.68) and −2.83 (±5.20), respectively. There was a robust correlation (Interclass Correlation Coefficient = 0.99) in evaluating the intra-observer reliability of the modified Friedman's method for measuring the glenoid version. The interobserver interclass correlation coefficient was moderate (ICC = 0.58).

Table 2.

Labral and glenoid versions of all 33 shoulders whose MRI studies were performed at our institution. SD, standard deviation; CI, Confidence Interval.

Sample characteristics	n	%	Mean (±SD)	Upper CI	Lower CI
Labral version	33	100	−3.98 (±5.68)	−1.96	−6.00
Anteversion	11	33.3	2.23 (±2.89)
Retroversion	22	66.7	−7.08 (±3.87)
Bony version	33	100	−2.83 (±5.20)	−0.98	−4.67
Anteversion	14	42.4	2.06 (±2.35)
Retroversion	19	57.6	−6.43 (±3.43)

While taking the 23 shoulders with a minimum one-year follow-up into consideration (Mean follow-up of 50.25 months; Range, 12–108 months), the labral version group had 9 (39.1 %) anteverted glenoids and 14 (60.9 %) retroverted glenoids. In the bony version group, 11 (47.8 %) of the shoulders had anteversion, whereas 12 (52.2 %) had retroversion (Table 3). The mean (±SD) angle of all the categories is mentioned in Table 3. 87.5 % of the anteverted labral version also had an anteverted bony version. There was a high strength of association (bias-corrected Cramer's V = 0.69) between the two, implying similar bony and labral versions in the shoulders.

Table 3.

Labral and bony version summaries of the 23 shoulders had a minimum one-year follow-up. SD, standard deviation.

Sample characteristics	n	%	Mean (±SD)
Labral version
Anteversion	9	39.1	2.22 (±3.11)
Retroversion	14	60.9	−7.52 (±4.20)
Bony version
Anteversion	11	47.8	2.06 (±2.49)
Retroversion	12	52.2	−6.47 (±3.86)

3.3. Clinical outcome and correlation with version

The mean (±SD) postoperative ASES, Oxford score, and SANE of all the participants (n = 23) were 84.04 (±14.14), 43.38 (±3.81), and 78.57 (±17.40), respectively (Table 4). There was no statistically significant clinical difference between the anteverted and retroverted labral version (p = 0.342 for ASES, p = 0.826 for Oxford score, and p = 0.401 for SANE). Similarly, there was no statistically significant clinical difference between the anteverted and retroverted bony version (p = 0.079 for ASES, p = 0.720 for Oxford score, and p = 0.073 for SANE).

Table 4.

The mean (±SD) postoperative ASES, Oxford, and SANE scores of labral and bony versions in 23 shoulders.

		ASES		Oxford		SANE	p-value
Labral version	Anteversion	87.49 (±12.38)	p = 0.342	43.88 (±3.36)	p = 0.826	81.88 (±19.07)	p = 0.401
	Retroversion	81.91 (±15.19)		43.08 (±4.17)		76.54 (±16.76)
Bony version	Anteversion	90.55 (±8.63)	p = 0.079	44.00 (±3.71)	p = 0.720	87.22 (±11.21)	p = 0.073
	Retroversion	79.15 (±15.77)		42.92 (±3.99)		72.08 (±18.76)

Traumatic vs. Atraumatic group: The mean labral version in the traumatic and atraumatic group was −6.99 ± 7.05 and −2.27 ± 5.29, respectively. The mean bony version in the traumatic and atraumatic group was −3.52 ± 7.17 and −1.90 ± 4.63, respectively. Though there is a notable difference in the means of both groups, it was not statistically significant (labral version: p = 0.193, bony version: p = 0.818). Further, the mean ASES of the atraumatic and traumatic group was 84.36 ± 14.79 and 86.42 ± 12.34, respectively. The mean SANE in the atraumatic and traumatic groups was 80 ± 17.51 and 80 ± 18.26, respectively. There were no statistically significant differences between the clinical outcomes of the atraumatic and traumatic groups (ASES: p = 0.92, SANE: p = 0.97).

Ability to throw at the final follow-up: After one year of follow-up with standard rehabilitation, all patients were asked if they could throw with the affected shoulder with preinjury velocity. Of the 23 shoulders, 10 patients (43.47 %) reported that they could not throw with the previous velocity and could not return to the original level of sports. Eight of those ten patients had retroverted glenoids, and two had anteverted glenoids.

4. Discussion

Our study concluded that patients with posterior instability have predominantly a retroverted glenoid version, bony and chondral. Nevertheless, despite having an observable difference in the glenoid version, the clinical outcomes after arthroscopic posterior shoulder stabilization showed no discernible differences. Further, nearly half of them (43.4 %) could not return to their original level of sports due to failure to achieve the ability to throw or smash.

Various methods have been utilized to assess the glenoid version. In 1986, Randelli et al. measured the glenoid version on the axial cuts of computed tomography scans of the scapula.¹⁶ Their glenoid version was defined on the angle subtended by, a line along the central axis of the scapula and another line connecting the anterior and posterior margins of the glenoid. The disadvantage of this method was that the scapular body index could not be accurately drawn on curved scapulae. Subsequently, in 1992, Friedman et al. proposed a Friedman's line, a reference line utilized to measure the glenoid version along the scapula axis.¹⁵ The Friedman method could bypass the hindrance offered by a curved scapula body. In 2010, Rouleau et al. compared the two methods (Randelli's and Freidman's) of measuring the glenoid version.¹⁷ The study showed excellent (>0.9) intraobserver reliability for both methods among all observers, which was consistent with our study (0.99). Nevertheless, the Friedman method yielded a marginally higher intraobserver correlation for the glenoid version. Rouleau et al. also determined the interobserver correlations, which were outstanding for both methods, with ICC values of 0.962 and 0.977 for the Randelli and Freidman methods, respectively.¹⁷ A third method of calculating the glenoid version developed by Poon and Ting uses the axis of the “glenoid vault” and the endosteal surface.¹⁸ This method has the advantage of mitigating the variable morphology and position of the glenoid body. Privitera et al. compared the Friedman and Poon and Ting methods; they found no statistical difference between the two methods (p = 0.20).¹⁹ We did not include the Poon and Ting method in our study because they had modified the definition of the glenoid version as an angle between the endosteal surface and the glenoid vault.

Lately, there has been a surge of interest in understanding the biomechanical effects of the glenoid version on glenohumeral stability, primarily due to the increasing trend of shoulder arthroplasty. A cadaveric study by Nyffeler et al. on shoulder prostheses concluded that glenoid retroversion is associated with increased posterior loading and posterior displacement of the glenoid, resulting in a more posteriorly directed joint force on the retroverted glenoid.²⁰ Imhoff et al. investigated the impact of the glenoid version on the centralization of the loaded humeral head and studied the versions of the glenoid in three states of the labrum-intact, detached, and repaired.²¹ They noted that retroversion, particularly when surpassing 10°, was a crucial factor contributing to the centralization and posterior translation of the humeral head onto the center of the glenoid.²¹ Additionally, they also emphasized that isolated posterior labral repair significantly prevents posterior translation.²¹

In a clinical setting, Gottschalk et al. evaluated 143 shoulders that underwent arthroscopic stabilization for both posterior and anterior labral tears; they observed that the shoulders with posterior labral tear had more degrees of retroversion when compared to the ones with anterior instability.²² Similarly, a prospective study spanning four years conducted by Owens et al. concluded that there is a higher degree of retroversion in athletes with posterior shoulder instability, concordant with our study. Furthermore, this study also demonstrated that with each 1° increase in retroversion, there was a 17 % greater probability of experiencing posterior shoulder instability in the future.² However, they did not evaluate any intervention or its clinical outcomes. In a prospective study conducted in 2006, Bradley et al. assessed 91 athletes (100 shoulders) with unidirectional posterior instability and compared them with 51 shoulders of contact athletes as control.⁷ Compared to the controls, patients with posterior instability had higher degrees of chondro- and osseous glenoid retroversion.⁷ Their conclusion also affirmed the effectiveness of posterior labral repair as a treatment for individuals with posterior shoulder instability. They also reported that only 67 % of patients could return to the previous level of sports. Another recent study in 2018, Bradley et al. concluded that the glenoid version does not affect clinical outcomes or risk of revision surgery.²³

In a study on patients with posteroinferior instability, Kim et al. categorized the shoulders of 31 patients into three types of labral lesions based on MRI arthrogram findings. “Type I as separation without displacement, type II as incomplete concealed separation, type III as loss of contour, and Type IV as flap tears.”²⁴ In the control group, the glenoid chondrolabral retroversion measured 2.3°, while the glenoid with types II and III posterior labral tear displayed a higher retroversion of 7.0°. Additionally, they observed no clinical distinction in outcomes between patients with and without chondrolabral glenoid retroversion, aligning with our findings. Christensen et al. noted that patients with increased glenoid bony version required surgical intervention, but this did not stand steadfast with the chondrolabral version.²⁵ In contrast, our study revealed a high association between the bony and the labral versions, indicating that the bony and labral versions of the same shoulder will generally be similar.

Mauro et al. prospectively studied 118 shoulders with unidirectional recurrent posterior shoulder instability, which was treated by arthroscopic posterior labral repair.²⁶ They assessed the glenoid version, glenoid width, and labral weight on MRI and measured three types of glenoid versions-bony, chondral, and labral and compared their pre- and postoperative ASES and pain scores. They reported superior ASES and lower pain scores in patients with narrower and more anteverted glenoid bone in comparison to those with broader and more retroverted glenoid. Furthermore, based on the glenoid bony version, Mauro et al. reported no significant differences in the clinical outcome scores, which was consistent with our study.

Currently, the bone of contention is whether the retroversion of the glenoid is a developmental state or a secondary phenomenon caused by repetitive humeral head subluxation.

Regarding difficulty in throwing ability after the surgery, ten shoulders (43.4 %) could not throw at one-year follow-up, of which five were involved in overhead sporting activity before the injury. Interestingly, most patients who could not throw after one year of posterior shoulder stabilization had retroverted glenoids. However, this result may be skewed as the prevalence of retroverted glenoids in our study is higher. In literature, return to the act of throwing has been studied only on athletes (mainly baseball players) based on return to sports. The throwing ability after posterior shoulder stabilization has not been evaluated in the non-sporting population. In a systemic review by Matar et al. the rate of successful return to sports following posterior shoulder stabilization ranged from 57.9 % to 100 %.²⁷ The high degree of variability may be because of the varied presentations and mechanisms, with some patients demonstrating frank posterior dislocation, others indicating only painful subluxation, and some just having vague shoulder pain. Due to this heterogeneity, comparing the return to sport or throwing activities can be rather challenging.

4.1. Limitation of our study

It comprises a small number of subjects, possibly because posterior instability is always less common than anterior and more ambiguous than its counterpart to establishing the clinicoradiological diagnosis. Secondly, the number of cases in our study that had posterior shoulder stabilization initially was higher, but this number dropped further due to the non-availability of MRI and loss in follow-up. Similar rates of dropout are reported in other studies, too.²³ Notwithstanding the limited sample size, our study concluded that glenoid retroversion constitutes a substantial risk factor for posterior shoulder instability. Third, though we did not have any control group, the comparative evidence from the literature is sufficient to prove that retroversion is more common in posterior instability than anterior.⁹ Fourth, we did not assess any bone loss in patients with posterior labrum tears, as none underwent a CT scan to determine bone loss. However, arthroscopic observation did not reveal any bone loss.

5. Conclusion

In conclusion, this study indicates a higher prevalence of retroverted glenoid versions and a strong association between labral and bony versions. Nevertheless, the glenoid version does not appear to affect the clinical outcome. However, further research is needed to validate these findings and explore potential treatment implications.

Financial remuneration to the authors

None.

Ethical approval

IRB approval and the approval number- IEC approval obtained - IEC/396/2022.

The study has been conducted in accordance with the ethical principles mentioned in the Declaration of Helsinski (2013).

Funding information

The study is not funded by institution or any other agency. Authors have not received any grant for the same.

Patient consent

The patients consent was obtained for the study.

CRediT authorship contribution statement

Sashitemjen Aier: Data curation, Formal analysis, primary Manuscript preparation. Bishak Reddy: Data curation, Formal analysis, primary Manuscript preparation. Vivek Pandey: Conceptualization, Methodology, Supervision, Validation, Writing – review & editing.

Declaration of competing interest

None.