Arm positions with increased risk of subscapularis external impingement at the subcoracoid arch

Article information

Clin Shoulder Elb. 2025;28(4):480-488

Publication date (electronic) : 2025 November 21

doi : https://doi.org/10.5397/cise.2025.00213

Su Cheol Kim¹, Michelle H. McGarry^2,#, Thay Q. Lee^2,#, Jae Chul Yoo^,1

¹Department of Orthopedic Surgery, Samsung Medical Center, Sungkyunkwan University School of Medicine, Seoul, Korea

²Orthopaedic Biomechanics Laboratory, Congress Medical Foundation, Pasadena, CA, USA

Correspondence to: Jae Chul Yoo Department of Orthopedic Surgery, Samsung Medical Center, Sungkyunkwan University School of Medicine, 81 Irwon-ro, Gangnam-gu, Seoul 06351, Korea Tel: +82-2-3410-3509, Fax: +82-2-3410-0061, E-mail: shoulderyoo@gmail.com

#Current affiliation: Orthopaedic Biomechanics Laboratory, Department of Orthopaedic Surgery, Creighton University School of Medicine, Omaha, NE, USA

Received 2025 March 4; Revised 2025 August 9; Accepted 2025 August 19.

Abstract

INTRODUCTION

Once designated the “forgotten tendon,” the subscapularis has been increasingly characterized through advances in shoulder arthroscopy [1,2]. Despite some variability, many studies have reported subscapularis tear classifications and patterns, enhancing the understanding of repair indications and techniques [3-5]. Nevertheless, the factors resulting in a subscapularis tendon tear have not been fully established [3,6-9]. Some reports suggest that the repetitive motion of the long head of the biceps (LHB) tendon may contribute to leading-edge damage to the tendon, whereas others have suggested tendon impingement against adjacent intra- or extra-articular structures [10,11].

Gerber et al. [12] reported that repetitive internal rotation (IR) combined with forward flexion (FF) can lead to subscapularis and LHB pulley damage owing to contact with the anterior glenoid labrum, which is described as internal impingement. However, other studies have demonstrated that specific arm motions can cause contact between the rotator cuff tendons and structures external to the glenohumeral joint, representing the primary mechanism of bursal-sided rotator cuff tears [13]. These structures, including the acromion, acromioclavicular joint, coracoacromial ligament, coracoid process, and conjoint tendon, are referred to as coracoacromial arches. The mechanism involved is commonly described as an external impingement [13,14].

Recent studies using various imaging modalities have reported the association between a short coracohumeral distance (CHD) and an increased risk of subscapularis tears [10,15,16]. Radiologic and clinical evidence supporting this includes findings from cases of large subscapularis tears, where ultrasound or magnetic resonance imaging (MRI) revealed a narrowed CHD and arthroscopy revealed degeneration of the conjoint tendon [16,17]. Nonetheless, it remains unclear whether subscapularis tears are caused by impingement or whether impingement is a consequence of a massive tear [17]. Furthermore, CHD measurements from imaging fail to capture the dynamic movement of the shoulder [12]. A cadaveric study with precise measurements of the contact area and pressure between the subscapularis and coracoacromial arch during glenohumeral motion would provide valuable insights into the characteristics of subcoracoid external impingement of the subscapularis tendon.

Therefore, this cadaveric biomechanical study aimed to identify arm positions that could increase the risk of subscapularis tears due to subcoracoid impingement. This study hypothesized that FF with IR, which mimics the Hawkins test position, results in high subcoracoid contact area and pressure.

METHODS

Institutional Review Board approval was waived by research institution (Congress Medical Foundation, Pasadena, CA, USA) for this basic science cadaveric study. Cadaveric specimens were obtained through an official body donation program and informed consent for research and educational use was obtained prior to death.

Measurement Setting

The scapula was fixed to a custom metal plate with three large bolts, and the plate was fixed to a custom-made shoulder testing system with an anterior tilt of 20° to simulate the resting in vivo position of the scapula. The transected humerus was fixed to an intramedullary rod and cylinder with six surrounding screws, and the humeral rod was attached to a custom device with a hollow-shaft angle potentiometer (Novotechnik US; precision, 0.5°), which was used to measure the humeral range of motion. The device was attached to an arc that enabled humeral axial rotation, abduction (ABD), and FF. The specimens were kept moist in normal saline to prevent dehydration during preparation and testing (Fig. 1A).

Fig. 1.

Specimen preparation and measurement settings. (A) The scapula was secured to a custom-made shoulder testing system with an anterior tilt of 20°. The transected humerus was fixed to an intramedullary rod, allowing movement in humeral axial rotation, abduction, and forward flexion. (B, C) The anterior and posterior views of the shoulder specimen show three fishing lines for subscapularis loading (arrowheads), three fishing lines for the infraspinatus and teres minor (arrows), and one line for the long head of the biceps tendon (asterisk).

The suture loop was connected to a Dacron fishing line (Izorline International), and the fishing lines were directed to mimic the loading of each rotator cuff and biceps brachii muscle. Adjustable pulleys and a positioning plate were used to approximate physiological muscle force vectors without friction. A muscle load of 10 N (2.2 lbf) was applied to each fishing line. Specifically, two lines were applied to the supraspinatus (20 N), three to the subscapularis (30 N), and three to the infraspinatus or teres minor (30 N). Additionally, one line was applied to the LHB tendon (10 N). These forces were determined from cross-sectional area measurements and electromyography studies (Fig. 1B and C) [18] .

A flexible pressure-mapping sensor (Tekscan, model 4201; saturation pressure, 13.8 MPa) was used to measure the contact area, force, and mean and peak pressure between the subcoracoid arch and subscapularis complex [19]. The subscapularis complex was defined as the upper insertion of the subscapularis tendon, LHB tendon, and pulley. The pressure-mapping sensor was approximately 0.1 mm thick, with a size of 8×20 mm. To ensure full coverage of the contact surface, each corner of the sensor was sutured along the coracoid and acromial attachments of the coracoacromial ligament, spanning the entire undersurface from the coracoid to the acromion. To minimize potential artifacts, the soft tissue overlying the subscapularis tendon and undersurfaces of the coracoacromial ligament, acromion, and coracoid were carefully removed. The sensors were calibrated using a two-point calibration method by applying 20 and 80 N loads using a material-testing machine (Instron, model 4411) (Fig. 2A).

Fig. 2.

(A) Installation of a pressure-mapping sensor under the subcoracoid arch. The size of the sensor was 8×20 mm, and each corner was sutured from the coracoid to the acromial attachment of the coracoacromial ligament. (B) The printout from the pressure-mapping sensor software shows red squares indicating areas of high contact pressure. The direction of this contact was located near the coracoid process.

RESULTS

In this study, the overall incidence of subcoracoid impingement was 52.1% (25/48; 95% CI, 37.8–66.4%) in the total testing conditions (Table 1). Impingement occurred most frequently when the shoulder was at 60° FF with maximal ER (83.3%), whereas no impingement was observed at 20° ABD with maximal IR. The incidence of subcoracoid impingement was significantly higher at 60° elevation than at 20° elevation (16/24 vs. 9/24, P=0.043). No significant differences in the incidence of subcoracoid impingement were observed between maximal ER and IR (14/24 vs. 11/24, P=0.386) or between ABD and FF (11/24 vs. 14/24, P=0.386). All contacts were observed on the coracoid side rather than the acromion side, which supported subscapularis impingement against the subcoracoid arch during arm motion (Fig. 2B).

Table 1.

Incidence of subcoracoid impingement (n=6)

The mean contact areas recorded by the sensors are shown in Fig. 3. The 20° FF with maximal IR exhibited the highest mean contact area, whereas the 60° FF with maximal IR exhibited a high contact area. Except for the 20° ABD with maximal IR, the one-way ANOVA showed significant differences in the mean contact area across the seven arm positions (P=0.009). Both 20° FF and 60° with maximal IR (20° FF, 318.9±155.2 mm²; 60° FF, 288.8±138.6 mm²) showed significantly higher contact area compared to both 20° FF and 60° with maximal ER, respectively (20° FF, 68.7±51.8 mm²; 60° FF, 122.2±91.0 mm²) (all P<0.02). Additionally, 60° ABD showed a high contact area (maximal ER, 230.1±136.0 mm²; maximal IR, 235.6±172.0 mm²). However, 60° ABD with maximal IR versus ER showed a similar contact area (P=0.948).

Fig. 3.

Mean contact area. Except for 20° abduction (ABD) with maximal internal rotation (IR), the mean contact area was significantly different across the seven arm testing positions (P=0.009). Both 20° and 60° forward flexion (FF) with maximal IR exhibited significantly larger contact areas compared to both 20° and 60° FF with maximal IR. Additionally, both 60° ABD with maximal IR and external rotation (ER) also showed high contact areas. Values are presented as mean±standard deviation. ^*P<0.05.

The mean contact pressures recorded by the sensors are shown in Fig. 4. Except for the 20° ABD with maximal IR, the one-way ANOVA showed no significant differences in mean contact pressure across the seven arm positions (P=0.188). The highest mean contact pressure was observed at 60° ABD with maximal ER (25.7±17.4 kPa). High mean contact pressure was also found with 20° FF with maximal IR (23.2±12.5 kPa), 60° FF with maximal IR (18.2±8.3 kPa), and 60° ABD with maximal IR (18.3±12.0 kPa), even though the difference was not significant.

Fig. 4.

Mean contact pressure. Except for 20° abduction (ABD) with maximal internal rotation (IR), there were no significant differences across the seven arm testing positions (P=0.188). Mean contact pressures exceeding 18 kPa were observed in the following positions: 60° ABD with maximal external rotation (ER) and IR, 20° forward flexion (FF) with maximal IR, and 60° FF with maximal IR. Values are presented as mean±standard deviation. ^*P<0.05.

The peak contact pressure recorded by the sensors is presented in Fig. 5, which shows a trend similar to that observed for the mean contact pressure. Except for 20° ABD with maximal IR, one-way ANOVA showed non-significant differences in the peak contact pressure across the seven arm positions (P=0.065). The highest peak contact pressure was observed at 20° FF with maximal IR (135.5±60.3 kPa). High peak contact pressure was also shown by 60° ABD with maximal ER (119.9±86.1 kPa), 60° FF with maximal IR (90.8±46.5 kPa), and 60° ABD with maximal IR (92.8±54.1 kPa), even though the difference was not significant.

Fig. 5.

Peak contact pressure. Except for 20° abduction (ABD) with maximal internal rotation (IR), there were no significant differences across the seven arm testing positions (P=0.065). Peak contact pressures exceeding 90 kPa were observed in the following positions: 60° ABD with maximal external rotation (ER) and IR, 20° forward flexion (FF) with maximal IR, and 60° FF with maximal IR. Values are presented as mean±standard deviation. ^*P<0.05.

The three arm positions with the greatest mean contact area, mean contact pressure, and peak contact pressure measured by the pressure-mapping sensor are summarized in Table 2. The positions of 20° FF with maximal IR, 60° ABD with maximal ER, and 60° ABD with maximal IR were consistently observed among the top three conditions across measurements.

Table 2.

Three arm positions with the greatest mean contact area, mean contact pressure, and peak contact pressure

DISCUSSION

In this cadaveric biomechanical study, external impingement of the subscapularis complex on the subcoracoid arch was investigated using a pressure-mapping sensor to reveal the arm position at risk for impingement. In general, high subcoracoid contact area and pressure were observed at 20° FF with maximal IR, 60° FF with maximal IR, and 60° ABD with maximal ER and IR arm positions, although not all differences were significant. Therefore, as we hypothesized, FF in the IR position was one of the arm positions at risk of subcoracoid impingement. In addition, ABD in the ER or IR position was also at risk of subcoracoid impingement.

The mechanism of subscapularis tears caused by impingement remains under investigation [15,21,22]. Previously, shoulder FF in the IR position was considered a risk factor for subscapularis tears. Gerber and Sebesta [23] reported that repetitive IR motion with an FF can cause subscapularis and LHB pulley damage by contacting the anterior glenoid labrum, which is known as internal impingement. Habermeyer et al. reported that adduction in the IR position can cause internal impingement of the articular-side subscapularis and the LHB pulley on arthroscopy [24].

External impingement has been reported in other biomechanical studies. Valadie et al. [14] performed a cadaveric cross-sectional study to investigate external impingement of the coracoacromial arch. They used nine cadaveric shoulders and analyzed them in Neer and Hawkins arm positions. They observed the impingement of rotator cuff tendons against the coracoacromial arch through cadaver dissection and direct visualization. Contrary to a previous claim, they reported a low incidence of external impingement in each arm position: 40% (2/5) in the Neer test position and 25% (1/4) in the Hawkins test position [14]. However, a key limitation of this study is its static design. Furthermore, one or a few slices of the cross-section cannot represent all the impingement results. The analysis of the impingement was also based on visualization, which did not provide an accurate contact measurement.

Other studies have attempted to assess external impingement using static imaging modalities such as computed tomography (CT) and MRI. Gerber et al. [12] investigated the subcoracoid external impingement of the rotator cuff using CT scans and analyzed 40 shoulders in the neutral position and 16 shoulders in FF with IR (Hawkins position). They observed that the CHD was reduced to a mean of 6.7 mm in the latter position, and that the prominent coracoid process contributed to the mechanical contact between the lesser tuberosity and coracoid in this provocative position.

Although the study performed a quantitative evaluation, it was limited by its static and non-dynamic nature and lacked a direct measurement of mechanical contact or pressure [12]. In contrast, our study used a pressure-mapping system to directly and dynamically measure the contact pressure between the subscapularis and surrounding structures during controlled motion, allowing more precise and functional subcoracoid impingement mechanics.

Dugarte et al. [25] conducted an anatomical study on 418 cadaveric shoulders to measure subcoracoid impingement and CHD under various conditions. They found that a narrower CHD was associated with shoulder IR and older female sex. However, the study had significant limitations. Subcoracoid impingement was assessed through direct visualization, and CHD was measured using a caliper, which could result in inaccuracy. Moreover, the absence of soft tissue, including the rotator cuff and joint capsule, limited the ability to replicate normal shoulder kinematics [25].

Unlike previous studies, our study used an accurate pressure-mapping sensor to determine the exact location and pressure of the contact surface between the subscapularis complex and subcoracoid arch. The results indicated that most of the contact area was located near the coracoid process, which supports coracoid impingement in subscapularis tears. Additionally, our study preserved soft tissues, including the joint capsule and rotator cuff tendons, which may minimally alter glenohumeral kinematics, in contrast to the study by Dugarte et al. [25].

Similar to the studies conducted by Gerber et al. [12] and Dugarte et al. [25], our study showed high subcoracoid impingement contact area and pressure of the shoulder FF in the IR position, which mimics the Hawkins test, although the values were not significant. Furthermore, 20° FF with IR and 60° FF with IR showed the first and second highest contact area, respectively. Notably, 20° FF with IR exhibited the highest contact pressure.

In our study, ABD with IR showed high contact area and pressure. The mean contact area, pressure, and peak contact pressure values at this position were the third highest in our analysis. CT analysis of CHD by Gerber et al. [12] also reported ABD in the IR position as one of the subcoracoid impingement positions. From their study, ABD with the IR position was one of the smallest CHD, and they claimed that ABD with the IR position was the most sensitive way to detect subcoracoid impingement.

Based on these findings, FF with IR and ABD with IR may be risk factors for subcoracoid impingement. In a clinical study supporting our findings, Park et al. performed a retrospective analysis of 23 patients with subcoracoid impingement who underwent arthroscopic coracoplasty [26]. They reported a significant increase in IR in the coracoplasty group, supporting the relationship between subcoracoid impingement and arm IR.

In our study, 60° ABD with ER also showed high contact area and pressure and corresponded to 90° ABD with ER when considering scapulohumeral rhythm. Neviaser and Neviaser [27] reported that 90° ABD with ER induced humeral head anterior translation, resulting in humeral head contact with the coracoid, which is consistent with our findings. Furthermore, ABD with ER may contribute to subscapularis tears by increasing tendon strain. For example, golfers exhibit peak subscapularis muscle activation during the acceleration phase of the swing at ABD with ER and are vulnerable to subscapularis tears [28]. Further studies evaluating the tendon strain across various arm positions are required.

Despite the limited number of cadavers and the lack of significance in some comparisons, our study identified arm FF with IR, ABD with ER, and ABD with IR as risk factors of external impingement of the subscapularis. Shoulder FF with IR can be observed during labor work and sports-related activities, such as painting or cleaning walls, car washing by hand, overhead serve during tennis, and the deceleration phase of throwing [29]. Shoulder ABD with ER position could be observed during sports activities, such as freestyle and butterfly stroke swimming, the top-to-acceleration phase of golf swings, and cocking to the acceleration phase of throwing [29]. Shoulder ABD in the IR position could be observed during the bench press exercise, which has been reported as a risky position for subscapularis partial tears [30]. According to the findings of our study, repetitive motion in this arm position may increase external impingement of the subscapularis, contributing to tears.

This study has several limitations. First, the study included a relatively small number of cadaveric specimens, each with different characteristics of sex, age, and height. Additionally, morphological variations may exist in structures such as the coracoid angle, length of the coracoid process, distance between the lesser tuberosity and coracoid process, acromion shape, morphology of the lesser tuberosity, and version of the proximal humerus [6]. Although anatomical variability is a critical factor, and documentation of the coracoid and lesser tuberosity morphology would have strengthened the analysis, this aspect was not thoroughly evaluated in the present study. However, to minimize potential bias, all procedures were performed by a single experienced surgeon, and only shoulders with anatomically normal structures and minimal degenerative changes were included.

Second, cadaveric studies could not reproduce in vivo properties. The scapular body was fixed to a custom testing system with a 20° anterior tilt, and there was no scapulothoracic motion during shoulder movement [31]. Given that dynamic scapular motion significantly influences the subcoracoid space, this represents a major limitation in a cadaveric study. In addition, although muscle loading was applied, the rotator cuff muscles were dissected from the scapula, and the periscapular muscles, such as the deltoid, latissimus dorsi, and teres major, were not preserved. This condition may have affected the physiological glenohumeral kinematics. Third, as the pressure-mapping sensor was sutured to the subcoracoid arch, the exact contact location on the humeral side could not be determined. Moreover, it remains unclear whether the detected contact is directly correlated with subscapularis tendon damage. Nonetheless, all contacts were observed on the coracoid side rather than on the acromial side, supporting the hypothesis of subscapularis impingement against the subcoracoid arch.

This study also presents several strengths. Previously, external impingement of the subscapular complex has been analyzed using fluoroscopy, CT or MRI, and cross-sectional cadaveric studies [12,15,23,32]. However, cadaveric experiments using accurate contact measurements with pressure-mapping sensors remain under-reported. Additionally, many surgeons have proposed mechanisms for subscapularis tears, but these are under investigation. This study may be helpful in revealing the mechanism of subscapularis tears related to external impingement. Its findings could contribute to the development of strategies for the prevention and repair of subscapularis tears.

CONCLUSIONS

This cadaveric study demonstrated that shoulders at 20° and 60° FF with IR and at 60° ABD with both ER and IR had increased contact with the subcoracoid arch. Although not all comparisons were significant, our findings suggest the potential for external subscapularis impingement in these positions.

Article information Continued