Acromion Morphology Is Associated With Failure of Arthroscopic Posterior Capsulolabral Repair

Abstract

Background:

Posterior shoulder instability has recently been linked to a higher and flatter acromion. However, the effect of acromial morphology on surgical outcomes has not been evaluated.

Purpose:

To evaluate differences in acromial morphology between patients undergoing primary and revision arthroscopic stabilization for posterior shoulder instability.

Study Design:

Case-control study; Level of evidence, 3.

Methods:

A series of patients who underwent either primary or revision arthroscopic stabilization for posterior shoulder instability between 2001 and 2022 were identified. Revision patients were matched to nonrevision patients based on age, sex, and sport. Comparisons of patient characteristics and acromial and glenoid morphology were done between groups using previously described magnetic resonance imaging parameters, including posterior acromial tilt (PAT), anterior acromial coverage (AAC), posterior acromial coverage (PAC), and posterior acromial height (PAH). Glenoid bone loss and glenoid version were also evaluated. Continuous variables were compared between groups using a 2-tailed Student t test.

Results:

A total of 37 patients who underwent revision posterior shoulder stabilization during the study period were identified and matched to 37 patients who underwent primary posterior shoulder stabilization. The mean ages of the primary and revision patients at the time of initial surgery were 21.5 ± 4.8 years (range, 13-35 years) and 19.6 ± 6.1 years (range, 12-45 years), respectively (P = .95). Men comprised 54% of each group. The mean PAT in the primary and revision groups was 51.6° and 53.7° (P = .32), while the mean PAH was 10 mm and 13.5 mm (P = .04), respectively. The mean AAC in the primary and revision groups was 4.44° and 7.10° (P = .20), while the mean PAC was 72.1° and 63.3°± 8.11° (P = .08), respectively. No difference was found in the glenoid version (6.22° in the primary group vs 8.08° in the revision group; P = .06). The mean glenoid bone loss was 0.57% (range, 0%-8.9%) in the primary group and 1.4% (range, 0%-5.8%) in the revision group (P = .02).

Conclusion:

A higher acromion was associated with requiring revision of previous arthroscopic posterior capsulolabral repair. PAT and coverage were not significantly different between patients undergoing primary versus revision repair. There was no significant difference in glenoid version between groups. Although the amount was small, glenoid bone loss was statistically higher in the revision group.

Posterior shoulder instability has recently been acknowledged as a common shoulder pathology that may be underdiagnosed.^1,13 Glenoid morphology and bone loss properties have been evaluated in the past, while interest in the effect of posterior acromial morphology has been of recent interest.^6,7,11,15 In 2019, Meyer et al ¹⁶ measured posterior acromial coverage (PAC), slope, and height on perfect scapular Y c-arm radiographs. They found that a more horizontally oriented acromial slope was associated with a higher risk of posterior instability. These authors concluded that a more vertical acromion may provide more bony restraint and prevent posterior instability. In a comparable study, Arner et al ⁵ in 2023 performed these same measures on typically obtained magnetic resonance imaging (MRI) studies and reported similar findings where a higher and flatter acromion was associated with posterior instability.

Despite the increased interest in posterior shoulder instability and acromial morphology, data regarding acromial morphology and its link to surgical failure and outcomes are unknown. Based on the findings of the aforementioned studies, the purpose of this study is to determine whether acromial morphology, as measured on MRI, is associated with higher arthroscopic capsulolabral repair failure. It was hypothesized that a higher and flatter posterior acromion would lead to a greater risk of failure of posterior shoulder capsulolabral repair.

Methods

After obtaining institutional review board approval, a consecutive series of 37 patients who failed posterior arthroscopic capsulolabral repair and underwent revision arthroscopic repair by a single surgeon (J.P.B.) was identified. Failure was defined as undergoing future revision surgery. These revision patients were then case matched by age, sex, and sport to 37 patients who underwent primary arthroscopic posterior capsulolabral repair and did not undergo revision surgery during the study period—from 2001 to 2021. Posterior instability was defined as patients with unidirectional posterior instability with a 2+ posterior load and shift, positive Kim and positive Jerk tests, and MRI findings consistent with the clinical picture. ⁴ All failed a trial of nonoperative treatment, including at least 6 weeks of physical therapy. The exclusion criteria were patients with multidirectional instability, voluntary dislocators, those with seizure disorders, or those with arthritis.

The follow-up was compared between the 2 groups. MRI measurements were done by 2 sports medicine fellowship-trained orthopaedic surgeons (J.W.A. and L.E.K.). These were done utilizing T2 sagittal sequences on a 3T MRI with 3 mm slices in a similar manner as previously described. ⁵ Both raters performed one round of measurements (raters 1 and 2, round 1) on both groups. One of the raters performed a second round of measurements (raters 1 or 2, round 2), with a minimum of 6 weeks between evaluations. These measures were posterior acromial tilt (PAT), anterior acromial coverage (AAC), PAC, and posterior acromial height (PAH). A reference line was first drawn from the inferior angle of the scapula through the center of the glenoid with extension superiorly. The PAT was then measured in degrees as the angle from the reference line to the most anterior and superior aspect of the acromion, to the most posteroinferior region (Figure 1). The AAC was then measured from the same reference line to the center of the glenoid, to the anterior-most aspect of the inferior acromion (Figure 2). The PAC was measured from the same starting point as the AAC to the posterior part of the inferior acromion (Figure 3). The PAH was then evaluated from the same reference line drawn to a line perpendicular to the reference line at the level of the most posterior inferior aspect of the acromion. The height from this line to the center of the glenoid was then measured (Figure 4). Glenoid version (degrees) and glenoid bone loss (percent) were measured as previously described. ⁹

Figure 1.

(A) A reference line is drawn on the sagittal T2 MRI sequence from the inferior angle of the scapula, through the center of the glenoid, with extension superiorly. (B) The PAT is then measured in degrees as the angle from the reference line to a line drawn from the most posterior and inferior portion of the acromion to the most anterior aspect. (Figures reprinted with permission from Arner JW, Nolte PC, Ruzbarsky JJ, et al. ⁵ Correlation of Acromial Morphology With Risk and Direction of Shoulder Instability: An MRI Study. The American Journal of Sports Medicine. 2023;51(12):3211-3216.). MRI, magnetic resonance imaging; PAT, posterior acromial tilt.

Figure 2.

A reference line is drawn on the sagittal T2 MRI sequence from the inferior angle of the scapula, through the center of the glenoid, with extension superiorly. The AAC is then measured from the point on the reference line at the center of the glenoid (Figure 1A) to the anterior-most aspect of the inferior acromion (degrees). (Figures reprinted with permission from Arner JW, Nolte PC, Ruzbarsky JJ, et al. ⁵ Correlation of Acromial Morphology With Risk and Direction of Shoulder Instability: An MRI Study. The American Journal of Sports Medicine. 2023;51(12):3211-3216.). AAC, anterior acromial coverage; MRI, magnetic resonance imaging.

Figure 3.

A reference line is drawn on the sagittal T2 MRI sequence from the inferior angle of the scapula, through the center of the glenoid, with extension superiorly. The PAC is then measured from the point on the reference line at the center of the glenoid (Figure 1A) to the most posterior part of the inferior acromion (degrees). (Figures reprinted with permission from Arner JW, Nolte PC, Ruzbarsky JJ, et al. ⁵ Correlation of Acromial Morphology With Risk and Direction of Shoulder Instability: An MRI Study. The American Journal of Sports Medicine. 2023;51(12):3211-3216.). PAC, posterior acromial coverage; MRI, magnetic resonance imaging.

Figure 4.

A reference line is drawn on the sagittal T2 MRI sequence from the inferior angle of the scapula, through the center of the glenoid, with extension superiorly. The PAH is then measured after drawing a line perpendicular to the reference line (Figure 1A) to the most posterior inferior aspect of the acromion. The distance from this line to the center of the glenoid is then measured in millimeters (orange bracket). (Figures reprinted with permission from Arner JW, Nolte PC, Ruzbarsky JJ, et al. ⁵ Correlation of Acromial Morphology With Risk and Direction of Shoulder Instability: An MRI Study. The American Journal of Sports Medicine. 2023;51(12):3211-3216.). MRI, magnetic resonance imaging; PAH, posterior acromial height.

A total of 37 athletes who underwent revision surgery by the same senior surgeon for posterior shoulder instability were matched by sex, age, and sport to 37 patients who underwent primary capsulolabral repair for posterior instability. Age matching was performed by matching each patient to their nearest neighbor without a global distance measure that was applicable to the total cohort. Continuous data were reported as means with standard deviations and 95% CIs. The intraclass correlation coefficient (ICC) was calculated for rounds 1 and 2 of a single rater to assess intrarater reliability, and for the first round of raters 1 and 2 to assess interrater reliability. Agreement was graded as described by Landis and Koch ¹⁴ (with ICC <0.20 corresponding to slight agreement, 0.21-0.40 to fair agreement, 0.41-0.60 to moderate agreement, 0.61-0.80 to substantial agreement, and >0.81 to almost perfect agreement).

Normality of all acromial measurements was assessed and confirmed using the Shapiro-Wilk test. Acromial data were compared using paired t tests, and 2-tailed P values were calculated with a significance level set at P < .05. The Mann-Whitney U test was used to compare age, with a significance level also set at P < .05. Statistical analysis was performed using Excel Version 2404(Microsoft) and SPSS Version 28.0.1(IBM).

Results

After matching and applying the inclusion and exclusion criteria, 37 patients (20 men [54%]; 17 women [46%]) were included in each group. The mean age of the primary group was 21.5 ± 4.8 years (range, 13-35 years), and the mean age of the revision group was 19.6 ± 6.1 years (range, 12-45 years) (P = .95). The primary group had a mean follow-up of 6.7 years (range, 2.3-14.3 years), while the revision group had a mean follow-up of 8.9 years (range, 2-21.3 years) (P = .141). The sports that were matched were as follows: 13 patients, football; 7 patients, softball; 3 patients, baseball; 3 patients, wrestling; 3 patients, swimming; 2 patients, basketball; 2 patients, weight lifting; 2 patients, gymnastics; 1 patients, rugby; and 1 patient, track.

The mean values for the measurements (PAC, PAH, PAT, and AAC) of the primary and revision groups are demonstrated in Table 1. Group comparisons revealed a statistically significant difference only in the measurements for the PAH, with revision patients demonstrating a higher PAH (P = .04). There was no significant difference between groups in PAC, PAT, or AAC measurements. No difference was found in glenoid version (6.22° in the primary group vs 8.08° in the revision group; P = .06). The mean glenoid bone loss was 0.57% (range, 0%-8.9%) in the primary group and 1.4% (range, 0%-5.8%) in the revision group (P = .02).

Table 1.

Comparison of the Mean Values of all Measurements Between the Primary and Revision Groups ^a

	Primary Group			Revision Group
	Mean	SD	95% CI	Mean	SD	95% CI	P
PAC, deg	72.1	10	52.1 to 92.1	68.3	8.1	52.1 to 84.5	.08
PAH, mm	10	7.2	−4.4 to 24.4	13.5	6.8	−0.1 to 27.1	.04
PAT, deg	51.6	8.6	34.4 to 68.8	53.7	10	33.7 to 73.7	.32
AAC, deg	4.4	7.3	−10.2 to 19	7.1	9.9	−12.7 to 26.9	.20

^aThe bold P value indicates statistical significance. AAC, anterior acromial coverage; PAC, posterior acromial coverage; PAH, posterior acromial height; PAT, posterior acromial tilt.

Inter- and intrarater reliability with ICCs and level of agreement are demonstrated in Table 2. The intrarater reliability for rounds 1 and 2 of a single rater was “substantial” to “almost perfect” and interrater reliability for round 1 of each rater was “almost perfect” according to Landis and Koch. ¹⁴

Table 2.

Intra- and Interrater Reliability, With ICC and Agreement According to Koch and Landis ^{14 a}

	ICC	Agreement
PAC, deg
Revision
Intrarater reliability	0.790	Substantial
Interrater reliability	0.883	Almost perfect
Primary
Intra-rater reliability	0.957	Almost perfect
Inter-rater reliability	0.978	Almost perfect
PAH, mm
Revision
Intrarater reliability	0.697	Substantial
Interrater reliability	0.821	Almost perfect
Primary
Intrarater reliability	0.961	Almost perfect
Interrater reliability	0.980	Almost perfect
PAT, deg
Revision
Intrarater reliability	0.923	Almost perfect
Interrater reliability	0.960	Almost perfect
Primary
Intrarater reliability	0.925	Almost perfect
Interrater reliability	0.961	Almost perfect
AAC, deg
Revision
Intrarater reliability	0.937	Almost perfect
Interrater reliability	0.968	Almost perfect
Primary
Intrarater reliability	0.891	Almost perfect
Interrater reliability	0.943	Almost perfect

^aAAC, anterior acromial coverage; ICC, intraclass correlation coefficient; PAC, posterior acromial coverage; PAH, posterior acromial height; PAT, posterior acromial tilt.

Discussion

The most important finding of this study is that failure of arthroscopic posterior shoulder capsulolabral repair is associated with a higher acromion (P = .04) (Figure 5). There is no association with failure of arthroscopic repair and PAT, PAC, or AAC. These results are valuable in surgeon and patient education regarding risk factors for failure of arthroscopic posterior shoulder capsulolabral repair. However, a difference of 3.5 mm in PAH is a small amount. Furthermore, acromial tilt is not associated with requiring revision surgery. Many factors are likely at play in these complex patients; nonetheless, in addition to glenoid morphology and patient factors, acromial morphology does appear to play a role in the success of arthroscopic posterior capsulolabral repair. Previous studies have demonstrated that a “higher and flatter” acromion may predispose patients to posterior shoulder instability, and a higher acromion may be associated with failure of arthroscopic repair.

Figure 5.

Example of a higher posterior acromion, which was associated with failure of arthroscopic posterior capsulolabral repair on CT scan. The present study evaluated measures on MRI. (Figures reprinted with permission from Arner JW, Nolte PC, Ruzbarsky JJ, et al. ⁵ Correlation of Acromial Morphology With Risk and Direction of Shoulder Instability: An MRI Study. The American Journal of Sports Medicine. 2023;51(12):3211-3216.) CT, computed tomography; MRI, magnetic-resonance imaging.

In recent years, there has been an increased knowledge of risk factors for failure of arthroscopic posterior capsulolabral repair. Glenoid retroversion has been shown to predispose a large military population to developing posterior instability, while other studies have demonstrated posterior glenoid bone loss to be the biggest risk factor for failure of arthroscopic repair: 11% bone loss resulted in a 10 times higher failure rate, while 15% bone loss resulted in a 25 times higher failure rate.^{2,3,6,8-10,17}

Glenoid version and bone loss were also evaluated to determine whether these were confounders when evaluating the effect of the posterior acromion. Glenoid version was found to be no different between the groups. Glenoid bone loss did show a significant difference between groups (P = .02). However, all glenoid bone loss measures were <9%, with the median being zero in both groups. The mean glenoid bone loss in the primary group was 0.57% (range, 0%-8.9%) and 1.4% (range, 0%-5.8%) in the revision group. The authors believe that these small glenoid changes— particularly <6% in the revision group—are less likely to contribute to failure of arthroscopic posterior labral repair. Previous work has shown that critical glenoid bone loss of >11% increases the risk of failure of surgery. The largest percentage of bone loss in the revision group was 5.8%. With the small mean (0.57% vs 1.4%), median (0%), and range of bone loss, and with most patients having no bone loss, the authors believe the differences found in this study are more likely due to acromial morphology. A small amount of glenoid bone loss in conjunction with a higher and flatter acromion may be synergistic risk factors. This is an area of future required study.

More recently, the influence of acromial morphology on posterior instability has been evaluated. Meyer et al ¹⁶ first introduced this concept in 2019, where they retrospectively assessed acromial morphology on perfect scapular Y fluoroscopic radiographs on 41 unidirectional posterior instability patients and age and sex matched them to 41 anterior instability patients as well as 53 controls. They found that a higher and flatter acromion was associated with the development of posterior shoulder instability. These same results were redemonstrated recently on MRI with very good to excellent interobserver reliability. ⁵ In this 2023 MRI study, the acromion was significantly higher in the posterior group when compared with the anterior and control groups, while the PAT was flatter. ⁵

The influence of posterior acromial morphology on outcomes of surgical repair was previously not evaluated. The present study found that PAH was the only measure associated with failure of capsulolabral repair requiring revision. The PAT did not appear to be linked with surgical failure. These findings are in line with a study by Arner et al ⁵ who evaluated the development of posterior instability, which found acromial height on MRI to be the most highly correlated measure. The exact mechanism of the influence of the posterior acromion is unclear. We postulate that a lower and more vertical acromion may provide some support in decreasing recurrent humeral head subluxations when a posterior-directed force is placed. This may be partially protective against the development of symptomatic posterior labral tearing and glenoid bone loss, and allow better surgical outcomes of posterior capsulolabral repair. Meyer et al ¹⁶ have proposed that without a lower and more vertical bony acromion, the posterior capsulolabral complex must provide higher levels of stability in isolation. In the present study, those who failed capsulolabral repair had a 3.5 mm higher acromion, which admittedly is quite small. However, intra- and interrater reliability were shown to be high in both this study and previous studies. The consistency of these results in combination with those by Meyer et al ¹⁶ and Arner et al ⁵ makes the height of the posterior acromion an important parameter when evaluating posterior instability.

Interventions to alter posterior acromial morphology are limited. The authors do not recommend surgical intervention to change acromial anatomy to decrease posterior shoulder instability. Posterior acromial osteotomy has been described; nonetheless, it is not presently done or recommended by the authors because of the small and thin bony characteristics and the known complications and high failure rates seen with open reduction and internal fixation of acromial stress fractures after reverse total shoulder arthroplasty.^12,21 Posterior acromial bone augmentation has been described, and a recent biomechanical cadaveric study has shown that this does successfully increase the force required to subluxate the humeral head posteriorly. ²⁰ Although limited clinical data exist, the authors believe that a bone block augmentation of the acromion is preferable to an osteotomy due to morbidity and risk of complications. ¹⁸ Currently, the authors are not performing either of these clinically.

Limitations

Limitations of the present study include the use of non-reformatted MRIs. However, the authors believe that utilizing MRIs without reformatting on the axis of the glenoid is likely more clinically applicable, as it is unlikely that special software would be widely used. Although acromial morphology measures appear to be an important piece of posterior shoulder instability, clinical utilization of measures, particularly with small differences, can be burdensome for clinicians to use routinely. A CT scan may allow more accurate measurements. However, evaluation on MRI is likely more clinically useful, as these studies are typically obtained on patients with instability. Further, measurement of glenoid bone loss has been shown to be accurate on MRI. ¹⁹ This study also did not evaluate other patient parameters, which also play an important role, such as size and height, or race. For example, large individuals may gravitate toward American football linemen, which may affect their acromial morphology and these results. A higher number of revision patients would add to the power of this study; nonetheless, revision arthroscopic posterior capsulolabral repair is relatively rare.

Conclusion

A higher acromion is associated with requiring revision of previous arthroscopic posterior capsulolabral repair. PAT and coverage are not significantly different between patients undergoing primary versus revision repair.