The role of scapular morphology in traumatic rotator cuff tears and greater tuberosity fractures: A retrospective study

Abstract

Background

Traumatic postero-superior Rotator Cuff Tears (RCT) and isolated Greater Tuberosity fractures (GTF) are equivalent injuries resulting in significant shoulder dysfunction if left retracted or displaced. The difference in morphometric aetiology is unclear. A raised critical shoulder angle (CSA) has been associated with rotator cuff degeneration. We hypothesised that traumatic RCT is associated with a raised CSA when compared to GTF.

Methods

A retrospective study was conducted across the two trauma units in our institution. All patients between the period of 2010 and 2020 with Traumatic GTF or RCT assessed on cross-sectional imaging (CT or MRI) were identified. Patients were case-matched by age, gender, mechanism and laterality of injury. The primary outcome measurement was the Critical Shoulder Angle (CSA). Other radiographic features of subacromial degenerative change, mechanism of injury, association with shoulder dislocation and delay to diagnosis were also compared.

Results

Eighty patients met the inclusion criteria(40 traumatic RCT and 40 GTF). The mean age was 61.8 years with 58(72.5%) left-sided injuries. Thirty-four (43%) were female.

The mean CSA was 3.96° higher in the RCT group (95% CI 2.5 to 5.41, p < 0.05). A CSA of 33.73 gave a sensitivity of 0.68 and a specificity of 0.8 to differentiate between RCT and GTF. Patients with RCT were far more likely to display subacromial degenerate changes and experience a significant delay in diagnosis, whereas those with GTF were more likely to have suffered a shoulder dislocation.

Conclusions

Patients with traumatic RCT have radiographic features and scapular morphology associated with degenerative rotator cuff disease compared to those with GTF. This supports the theory that tears occur on the background of pre-existing tendon degeneration. Careful assessment of these parameters, combined with clinical assessment, may help guide the provision of appropriate diagnostic imaging.

Level of evidence

III.

Highlights

• •Pre-existing rotator cuff tendon degeneration predisposes to tendon tears.
• •Greater Tuberosity fractures are more associated with shoulder dislocation than rotator cuff tears.
• •Rotator cuff tears are associated with a higher critical shoulder angle than those with a greater tuberosity fracture.

Abbreviations

RCT

Rotator Cuff Tear

GTF

Greater Tuberosity Fracture

CSA

Critical Shoulder Angle

ICC

Interrater Correlation Coefficient

1. Introduction

Rotator cuff tears (RCTs)1 and isolated fractures of the Greater Tuberosity (GTFs)2 are common traumatic sequelae. They often occur with anterior shoulder dislocation or direct trauma. Missed injuries remain a persistent problem, from a clinical and medico-legal perspective.¹ Traumatic RCT can be challenging to diagnose as they cannot be directly visualised on plain radiographs.² Suspected RCT usually require an Ultrasound (US) or Magnetic Resonance Imaging (MRI).³ Similarly, GTF are often subtle and may not be detected on initial radiographs.⁴ Displaced tuberosity fractures are typically reduced and managed surgically. Non-displaced fractures are treated conservatively however these are closely monitored for secondary displacement.⁴ Despite the similar mechanism of injury, in some patients, the GTF remains intact and patients suffer an RCT instead. The evidence is not clear why some patients avulse their GT and others tear their rotator cuffs without avulsing their GTF despite the similar mechanism of injury and presentation.

Current guidelines recommend performing ultrasound in all patients above the age of 40 presenting with a shoulder dislocation to identify any associated cuff tear.³ However, most traumatic RCT patients do not present with a shoulder dislocation.⁵ Larger RCTs are a significant source of shoulder dysfunction, especially when treatment is delayed and tendon retraction occurs.⁶ Hence there is a clear need to understand the mechanics and aetiology that govern these common shoulder trauma pathologies to aid diagnostic and treatment pathways.

A possible explanation is the intrinsic patient anatomy and morphometric differences that may predispose to pre-existent rotator cuff degeneration. Multiple subacromial radiographic features have been associated with degenerative rotator cuff pathologies, but these are often difficult to define, do not discriminate between tears and tendinopathy, and suffer from significant variation of interobserver reliability.⁷ Recently, the association between coronal scapular morphology and different shoulder pathologies have been discovered.⁸ In 2013, Moor et al. introduced the concept of critical shoulder angle (CSA)3 as a reliable parameter to differentiate between primary osteoarthritis and degenerative rotator cuff tears.⁹ The CSA is the angle between the glenoid surface and a line connecting the inferior rim of the glenoid and the lateral tip of the acromion measured in the axial plane.¹⁰ A higher CSA was associated with degenerative RCT and these have since been replicated in multiple other studies. However, there is a paucity in the literature in assessing the association of CSA with traumatic shoulder injuries.

We hypothesised that patients with a traumatic RCT would have a higher CSA than patients who sustained an isolated GTF and have associated radiographic features of pre-existing degeneration.

2. Materials and methods

2.1. Study design

A retrospective study was conducted at two Trauma Units within our institution. All patients between 2010 and 2020 with isolated Traumatic GTF or full-thickness traumatic RCT were identified using clinical coding. Patients that met the inclusion criteria were then identified. The primary outcome was the CSA as measured on cross-sectional imaging (MRI or CT scan). The other primary outcome was any tuberosity and acromial radiographic features on a plain Anteroposterior (AP) shoulder radiograph. Secondary outcomes included the presence of shoulder dislocation.

HRA and local ethical approval was gained (IRAS ID 288623).

2.2. Patient selection

Table 1 shows the inclusion and exclusion criteria.

Table 1.

Inclusion and exclusion criteria of patient selection.

Inclusion Criteria	Exclusion Criteria
•Patients aged 40 to 80	•Missing demographic data
•Patients with clinically and radiographically confirmed: full-thickness traumatic RCT, Isolated GTF	•Unknown mechanism of injury
•Available cross-sectional imaging(Computed Tomography or Magnetic Resonance Imaging)	•Pre-existing shoulder pain or previous attendance for a shoulder problem

2.3. Data extraction

The groups were case-matched by age, laterality, gender and mechanism of injury. Age and mechanism of injury were chosen as they are the only known demographic parameters associated with full-thickness rotator cuff tears.¹¹ Mechanism of Injury was categorised as per the classifications of the Trauma Audit and Research Network.¹²

Radiographic data extracted included any subacromial morphological changes identified by the senior author (EI) on plain imaging, such as acromial sclerosis or spurring or greater tuberosity surface irregularity, intra-cortical cysts, sclerosis or spurring (RCT group only). The waiting time between the initial injury and diagnosis on cross-sectional imaging was also recorded.

2.4. Critical shoulder angle measurement

Critical shoulder angles were measured on shoulder coronal T1 weighted MRI (RCT cohort) and shoulder coronal CT (GTF cohort) that had been pre-formatted according to defined parameters according to the plane of the scapula body.¹³ The CSA was measured by an angle subtended between a line connecting the most inferior and superior borders of the bony glenoid fossa (corresponding to the mid-glenoid in the axial plane) and another line connecting the inferior border of the glenoid with the most inferolateral border of the acromion.⁹ All CSA angles for each patient were measured by 3 raters (EI, OM and KS). Raters were blinded to other raters’ measurements. The Interrater Correlation Coefficient (ICC) was calculated, and the average CSA reading was taken.

2.5. Statistical analysis

Tests of normality were assessed using the Shapiro-Wilk test. Inter-rater reliability of critical shoulder angle measurement was analysed using Interrater Correlation Coefficient (ICC).

The “Time to cross-sectional imaging” difference between the two groups was calculated using the Student's T-test.

Receiver operating characteristic (ROC) analyses was performed to assess the sensitivity and specificity of the CSA in determining the difference between RCT and GTF. Liu's cut off point estimation test was used to estimate the optimal CSA cut off point. Statistical significance will be set at <0.05. STATA™ was used for the statistical analysis.

3. Results

3.1. Demographic data

Patient selection was conducted using similar methodologies to other CSA studies. In reverse chronological order of scan date, 40 patients with traumatic GTF were identified out of all the patients with traumatic GTF presenting to two Trauma Units that met the inclusion criteria. The 40 patients with GTF were then case-matched with all the patients with traumatic RCT to find 40 equal matched traumatic RCT. Case matching was performed in reverse chronological order of scan date using patient age, laterality, gender and mechanism of injury. These variables were selected based on literature recommendations.¹⁴ Overall, the average age was 61 years in both groups. The most common mechanism of injury was a fall from less than 2 m. Forty one percent of patients had a shoulder dislocation in the overall study. The GTF cohort was associated with a higher rate of dislocations than the RCT (65% vs 18%) in the respective groups (Table 2).

Table 2.

Demographic data of patients meeting inclusion criteria.

	Traumatic Full-Thickness Rotator Cuff Tear (n=40)	Greater Tuberosity Fracture (n=40)
Mean Age(SD)	61.7(6.7)	61.8 (8.8)
Gender
•Male	23	23
•Female	17	17
Laterality
•Right	11	11
•Left	29	29
Mechanism of Injury
•Fall<2m	28	28
•Fall>2m	2	3
•RTA	4	3
•Sports Related	1	2
•Other	5	4
Shoulder Dislocation
•Yes	7	26
•No	33	14

3.2. Critical shoulder angle and radiographic findings

All RCT cohort patients had some features of subacromial degenerative disease on plain radiograph. 32(82%) had acromial changes (sclerosis or spurring) and 38(95%) had tuberosity changes (spurs, cysts, surface irregularity or sclerosis (Table 3). Due to presence of a fracture, it was not possible to comment on tuberosity changes in the GTF group. However, only 3(8%) of the GTF group had radiological acromial changes (p < 0.05).

Table 3.

Table showing radiographic findings and CSA results.

	Traumatic Full-Thickness Rotator Cuff Tear (n=40)	Greater Tuberosity Fracture (n=40)
Mean Time from presentation to MRI/CT(SD)	88.2 Days (SD 92.3)	9.8 Days (SD 10.7)
Mean CSA(Range, SD)	35.4° (30–46, SD 3.56)	31.4° (24.6–36.4, SD 2.94)
Acromial Changes
•Sclerosis	25	2
•Spurs	28	0
Tuberosity Changes
•Surface Irregularity	8	N/A
•Sclerosis	5	N/A
•Spurs	16	N/A
•Cysts	25	N/A

The inter-rater reliability coefficient was 0.81(ICC two-way mixed-effects model), indicating a high level of reliability in CSA measurement. The distribution of CSA in both cohorts is shown in Fig. 1. The mean CSA was significantly higher in the RCT group (3.96°; 95% CI 2.5 to 5.41, p < 0.05) (Table 3). The Receiver Operator Curve is shown in Fig. 2. The Area under the curve was 0.8 indicating high reliability. A CSA of 33.73° gives a sensitivity of 0.68 and a specificity of 0.8 to differentiate between traumatic RCT and GTF.

Fig. 1.

Box-and-whisker plot comparing RCT and GTF CSA values.

Fig. 2.

The Receiver Operator Curve gives an area under the curve of 0.8 (95% CI 0.71–0.9) indicating high reliability. The reference line is indicated in grey.

The mean time from presentation to CT in the GTF group was 9.8 days (SD 10.6). The mean time from presentation to MRI in the RCT cohort was 88.2 days(SD 16.1). The overall difference in presentation to further scan between the 2 groups was 78.3 days(P < 0.05, 95% CI 48.7–108).

4. Discussion

The study found that patients with a traumatic rotator cuff tear have a higher mean CSA and are more likely to have degenerative subacromial radiographic changes than those with an isolated greater tuberosity avulsion fracture. This supports the theory that traumatic RCT tend to occur in the presence of pre-existing tendon degeneration whereas GTF occurs through the avulsion of bone by a healthy tendon.

The mean CSA from our study was similar to other studies. Seo et al. found a mean critical shoulder angle of 34.3° in full-thickness tears, and other studies have reported a comparable CSA to our study (35.4°) in a degenerative RCT cohort. Whilst they measured the CSA on AP radiographs and not cross-sectional scans, there are published studies that have shown there is little difference in CSA measurements between the different imaging modalities.¹⁵ Similarly, a recent meta-analysis demonstrated that radiographic quality is a source of heterogeneity in studies that investigate the link between CSA and RCT and glenohumeral osteoarthritis.¹⁶ There is no published study to compare CSA measurement in GTF.

The radiological findings in the rotator cuff group show that there is a presence of acromial and tuberosity degeneration prior to the trauma. There is evidence that the presence of acromial spurs is associated with a high incidence of degenerative full-thickness cuff tears.¹⁷ Radiological signs that correlate well with rotator cuff arthropathy include acromial sclerosis, spurs, GT irregularity, and GT cysts.^18,19 In our study, there was a marked difference in subacromial radiological findings between RCT and GTF cohorts. The plain radiographs of the traumatic RCT cohort had a high presence of such features (Table 3). This further supports the assertion that patients who had a traumatic RCT tear have pre-existing features of rotator cuff degeneration prior to the trauma. Hence, traumatic RCT may be an acute-on-chronic process. This is similar to findings in other tendon pathologies such as Achilles tendon tears.

In Achilles tears, biopsies retrieved at surgery demonstrated chronic degenerative changes in ruptured Achilles tendons.²⁰ Whilst Achilles tendon injuries are extra-synovial and rotator cuff injuries are intra-synovial and do not undergo spontaneous healing, there may be a more similar immunopathological process between them than previously suggested. To our knowledge, there are no animal studies that have addressed this in other commonly injured tendons such as patella or tibialis posterior tendon.²¹ Our study findings reinforce the scientific rationale that tendon ruptures may be categorised as acute trauma of chronically degenerated tendons.

Early diagnosis of acute RCTs is helpful to assess the size and position of tears and guide treatment decisions to maximise clinical outcomes.⁵ In this study there was a significant delay to MRI in the RCT cohort with an average of 88.9 days from presentation (Table 3). This concern does not appear confined to our institution. Before the implementation of an effective new service pathway, a service evaluation of 30 rotator cuff patients from a large UK hospital found there was a mean delay of referral to treatment of 115 days following trauma.²² This has significant implications in patient outcomes as delayed surgical repair is often more challenging due to tendon retraction and degeneration and there is growing evidence for the early repair in appropriately selected patients. Comparatively the GTF cohort had a significantly quicker time to cross-sectional image with an average wait of 9.8 days to CT (Table 3). This is likely because fractures are more visible on plain radiographs and only the most obvious potentially displaced fracture would require a CT. However, the clinical patient pathways appear to be significantly different in both cohorts despite similar mechanisms of injury and potential sequelae.

Larger CSAs (>35°) have been proposed to result in a more superior deltoid muscle force vector due to the lateral projection of the acromion. The altered deltoid vectors result in increased superior shear forces on the rotator cuff muscles. This loading may be a risk factor for the development of rotator cuff tears.²³ Possibly, the majority of patients in our acute traumatic RCT cohort may have historically been exposed to higher repetitive rotator cuff loading before the trauma. Biomechanical studies have shown that the supraspinatus muscle must be activated to a greater extent to stabilise the shoulder in those with a high CSA.^24,25 This further explains the increased observation of degenerative rotator cuff tears in association with a large CSA due to the chronic abnormal loading patterns of the supraspinatus tendon. However, not every patient with an RCT had a high CSA and therefore other unknown biomechanical or biological factors may play a role in these patients. The CSA in the GTF group was comparable to the values seen in normal shoulders and associated with a higher rate of anterior shoulder dislocation. Despite the similarities in the mechanism of injury, this suggests that higher energy trauma was required to cause a bony avulsion rather than the presence of osteoporotic bone. Current guidelines favour an early US approach in patients who have suffered a shoulder dislocation to address rotator cuff integrity.³ However, patients with a CSA <33° who have suffered an anterior dislocation may instead have a GTF. The US has little diagnostic value in assessing the greater tuberosity.⁴ Therefore, when a fracture is not obvious and the CSA is < 33°, an MRI may be advantageous in diagnosing either, or an alternative diagnosis entirely.

Based on this study, we recommend radiology protocols are used to standardise the use of ‘true’ Grashey AP views on which the CSA can be measured. Clinicians should carefully interrogate plain radiographs, in conjunction with clinical assessment, to aid decision making in the diagnosis of RCT. Following shoulder trauma, the presence of a CSA >33° or subacromial degenerative changes should mandate further imaging to exclude RCT.

4.1. Limitations

This is a retrospective study and hence there is an element of selection bias however our patients were case-matched based on known prognostic factors. Ultrasound-only confirmed rotator cuff injuries were not included as it is not possible to measure the CSA. This may have resulted in selection bias of those tears that were less clinically apparent prompting the clinician to arrange an MRI, or felt to be more extensive requiring more detailed surgical planning. Our study was not blinded due to the nature of measuring CSA. As the vast majority of plain radiographs of the injured shoulder were not true AP views, we used cross-sectional imaging to calculate the CSA.

5. Conclusions

Patients with a traumatic rotator cuff tear are less likely to have suffered a dislocation, and more likely to have a higher critical shoulder angle with signs of subacromial degeneration than those with a greater tuberosity fracture. This supports a theory of pre-existing tendon degeneration that predisposes to tendon tears and has implications for greater streamlined diagnostic pathways.

Funding/sponsorship

This research did not receive any specific grant from funding agencies in the public, commercial or not-for-profit sectors.

Institutional ethical committee approval (for all human studies)

Approved IRAS ID 288623.

Author contributions

OM was involved in design, data collection, analysis, drafting of manuscript. KdeS was involved in data collection, drafting and review of manuscript. EI was involved in data collection, design and review of manuscript.

Declaration of competing interest

None.