Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2018 Oct 1.
Published in final edited form as: J Orthop Res. 2017 Mar 27;35(10):2329–2337. doi: 10.1002/jor.23515

Effect of glenohumeral elevation on subacromial supraspinatus compression risk during simulated reaching

Rebekah L Lawrence 1, Dustin M Schlangen 1, Katelyn A Schneider 1, Jonathan Schoenecker 1, Andrea L Senger 1, William C Starr 1, Justin L Staker 1, Jutta M Ellermann 2, Jonathan P Braman 3, Paula M Ludewig 1
PMCID: PMC5503805  NIHMSID: NIHMS843465  PMID: 28071815

Abstract

Mechanical subacromial rotator cuff compression is one theoretical mechanism in the pathogenesis of rotator cuff disease. However, the relationship between shoulder kinematics and mechanical subacromial rotator cuff compression across the range of humeral elevation motion is not well understood. The purpose of this study was to investigate the effect of humeral elevation on subacromial compression risk of the supraspinatus during a simulated functional reaching task. Three-dimensional anatomical models were reconstructed from shoulder magnetic resonance images acquired from 20 subjects (10 asymptomatic, 10 symptomatic). Standardized glenohumeral kinematics from a simulated reaching task were imposed on the anatomic models and analyzed at 0°, 30°, 60°, and 90° humerothoracic elevation. Five magnitudes of humeral retroversion were also imposed on the models at each angle of humerothoracic elevation to investigate the impact of retroversion on subacromial proximities. The minimum distance between the coracoacromial arch and supraspinatus tendon and footprint were quantified. When contact occurred, the magnitude of the intersecting volume between the supraspinatus tendon and coracoacromial arch was also quantified. The smallest minimum distance from the coracoacromial arch to the supraspinatus footprint occurred between 30–90°, while the smallest minimum distance to the supraspinatus tendon occurred between 0–60°. The magnitude of humeral retroversion did not significantly affect minimum distance to the supraspinatus tendon except at 60° or 90° humerothoracic elevation. The results of this study provide support for mechanical rotator cuff compression as a potential mechanism for the development of rotator cuff disease.

Keywords: subacromial impingement, supraspinatus, kinematics, shoulder

Introduction

Shoulder pain is the second most common cause of musculoskeletal symptoms1 and can substantially impact an individual’s ability to perform functional and sport-related activities.2 The most prevalent finding on magnetic resonance images in individuals with shoulder pain are abnormalities of the rotator cuff.3 While the etiology of rotator cuff disease is likely multi-factorial,4, 5 mechanical compression of the rotator cuff tendons during shoulder motion is one mechanism that has been commonly theorized. Mechanical subacromial rotator cuff compression, often termed subacromial impingement, is defined as compression and abrasion of the bursal side of the rotator cuff against the undersurface of the coracoacromial (CA) arch.6 The supraspinatus is especially vulnerable due to its location immediately beneath the CA arch and has a higher incidence of tearing compared to other rotator cuff muscles.3, 7 Importantly, mechanical subacromial rotator cuff compression should be considered as a potential extrinsic mechanism of rotator cuff disease distinct from the clinical diagnosis of “impingement syndrome” as this clinical label is widely agreed to be too broad and may include numerous pathoanatomic conditions.8–11

Subacromial compression of the rotator cuff tendons is generally believed to occur during shoulder elevation motion. However, the specific shoulder positions that place the rotator cuff at highest risk for compression remain unclear. Studies seeking to quantify subacromial rotator cuff compression across elevation ranges of motion have generally used bone-to-bone minimal distance measures between the humerus and acromion (i.e. acromiohumeral distance). These studies found the minimum acromiohumeral distance occurs near or above 90° humerothoracic elevation.12–18 This was initially interpreted to suggest the rotator cuff is at greatest risk for compression in this mid-range of arm elevation. However, acromiohumeral distance can be misleading since it is assessed independent of the location of the rotator cuff tendons. For example, Giphart et al. found the smallest acromiohumeral distance occurred at 83° humerothoracic elevation during scapular plane abduction.15 Although this study used fluoroscopy and could not specifically visualize the rotator cuff tendons, the authors identified when the minimum acromiohumeral distance vector was within the superior facet of the greater tuberosity and therefore compression of the supraspinatus footprint could have occurred. Using this approach, the authors found the supraspinatus footprint had passed under the acromion and was no longer at risk for compression by 72° humerothoracic elevation, even though the acromiohumeral distance was not yet at its minima. Consequently, traditional acromiohumeral distances measures may have limited significance relative to rotator cuff compression risk.

Despite the limitations of the measurement, studies utilizing acromiohumeral distance measures have helped refine our understanding of mechanical rotator cuff compression from the original descriptions by Meyer19 and Neer6. However, many questions remain affecting our ability to prevent, diagnose, and treat rotator cuff disease. Previous studies do not directly address the fundamental question of how subacromial proximities are affected by kinematic changes across a range of motion. Doing so will allow us to identify shoulder elevation positions that place the rotator cuff at highest risk for subacromial compression.

To date, no in vivo study of humeral elevation has examined three-dimensional minimum distances directly to the rotator cuff tendons or included the CA ligament as a potential site of compression. Furthermore, minimum distance measures only quantify risk for a single point on the surface of the tendon that does not necessarily capture the complexity of the phenomenon. When subacromial contact occurs, metrics that relate to compression or deformation risk may provide a more comprehensive assessment. The purpose of this study was to investigate the effect of humeral elevation (from 0–150°) on subacromial compression risk of the supraspinatus tendon during a simulated functional reaching task.

Methods

Subjects

Ten asymptomatic subjects and ten symptomatic subjects were recruited as a sample of convenience from other investigations or the university setting (TABLE 1). Symptomatic subjects were included if they had signs and symptoms consistent with the broad clinical diagnosis of “impingement syndrome”20 and were excluded if: 1) shoulder symptoms began following trauma, 2) symptoms were reproduced during cervical spine screening, or 3) if the subject had a history of shoulder fracture, dislocation, or inflammatory joint disease. Asymptomatic subjects were included if they had no previous history of shoulder pain. Both symptomatic and asymptomatic subjects were included to study positional compression risk across the population. The Institutional Review Board of the University of Minnesota approved the study protocol and subjects provided written informed consent prior to data collection.

TABLE 1.

Demographic Data

Asymptomatic (n=10) Symptomatic (n=10) p-value
Age 38.5 ± 12.8 43.0 ± 11.8 0.42
Height (cm) 172.8 ± 8.8 169.7 ± 9.9 0.46
Weight (kg) 75.9 ± 14.2 77.1 ± 16.5 0.86
Gender (male) 5 (50%) 4 (40%) 1.0
Side tested (dominant) 8 (80%) 10 (100%) 0.47
Open in a new tab

Values are mean ± SD unless otherwise indicated.

Data Processing

Magnetic resonance (MR) images of each subject’s symptomatic or dominant shoulder were obtained using a Siemens Magnetom SKYRA system at a magnetic field strength of 3 Tesla. Images were acquired utilizing a 3D gradient recalled echo sequence T1-VIBE with water excitation. The following acquisition parameters were applied: acquisition matrix 256 × 256, FOV 160 × 160 mm, slice thickness 0.6 mm, repetition time (TR) 10.9 ms, and echo time (TE) 4.8 ms, echo train length 1, number of averages 1, and flip angle 10. Digital image files (DICOMs) of the MR scans were imported into Mimics software (Materialise; Leuven, Belgium) where masks of the humerus, scapula, CA ligament, and supraspinatus were created. This extensive process involved methods of image segmentation including initial thresholding of voxel grey values followed by manual segmentation of the anatomical structures. The final masks of each structure were reviewed by a fellowship-trained shoulder surgeon before being rendered into subject-specific three-dimensional anatomical models (FIGURE 1).

FIGURE 1.

FIGURE 1

Open in a new tab

Example of a subject-specific anatomical model reconstructed from MR images. The footprint was defined as the portion of the supraspinatus tendon lateral to the articular margin of the humeral head and is denoted in black. The tendon was defined as the footprint and 1 cm medial (dark grey area), which has been described as the “critical zone”.22

Once the 3D models were created, the footprint and tendon areas of the supraspinatus were identified (FIGURE 1). The footprint area was defined as the portion of the muscle lateral to the articular margin on the humeral head.21 The tendon area of the supraspinatus was defined as the footprint area and 1 cm medial, which encompasses the “critical zone”22 where most supraspinatus tears originate. Although the supraspinatus tendon contains the footprint region, investigating proximity measures to the footprint alone allowed for direct comparison to previous studies12, 13, 15 that identified the footprints to help interpret acromiohumeral distances. Finally, the 3D models of the humerus, CA ligament, acromion process, and supraspinatus tendon and footprint were exported from Mimics as stereolithography (STL) files for further processing.

Humeral and scapular local coordinate systems were defined by digitizing anatomic landmarks on the 3D models within the Mimics global coordinate system. Coordinate systems were constructed based on the recommendations of the International Society of Biomechanics.23 However, due to the restricted field of view of the MR scanner, not all recommended anatomic landmarks were visible for digitization. Surrogate points were defined and confirmed through pilot testing to minimally impact axis alignment (SUPPLEMENTARY APPENDIX). Following construction of the coordinate systems, the humeral head was centered on the glenoid using a sphere fit to the articular surface of the humeral head and a circle fit to the inferior margin of the glenoid using least-squares algorithms.24 The glenohumeral joint was initially aligned to a neutral position such that the axes of the humeral coordinate system were parallel to the axes of the scapular coordinate system (i.e. all initial glenohumeral angles were 0°). The amount of subject-specific humeral retroversion could not be directly accounted for because the humeral epicondyles were outside the MR scanner’s field of view. Therefore, each subject’s humeral head was internally rotated relative to the scapula from the neutral position by 57.2° to account for the average humeral retroversion observed using a biceps groove measure.25

The aim of the study was to compare the effect of typical angular humeral elevation movement patterns on subacromial proximities. We utilized a modeling approach where, rather than measuring subject-specific kinematics at each of these angles, we applied standardized kinematics derived from an average data set from healthy subjects. With this approach, our primary interest is the effect of glenohumeral elevation (between-condition factor) on subacromial proximity. We therefore experimentally controlled the additional between-subject variance in kinematics at a given elevation/retroversion angle.

The standardized shoulder positions were defined as the average glenohumeral joint kinematics from a previous investigation that used bone-fixed electromagnetic sensors to quantify shoulder complex kinematics in asymptomatic individuals.26 Humerothoracic and glenohumeral data were collected during an unconstrained functional reaching task using anatomical coordinate systems defined according to published recommendations.23, 27 Humerothoracic position was described as plane of elevation (Y-axis), elevation (X-axis), and axial rotation (Y-axis) using an Y,X′,Y″ rotation sequence, and glenohumeral position was described as elevation (X-axis), plane of elevation (Z-axis), and axial rotation (Y-axis) using an X,Z′,Y″ rotation sequence.28 Mean values for glenohumeral elevation, plane of elevation, and axial rotation were calculated at 5° increments of humerothoracic elevation between 0° and 150° (TABLE S-1).26 A custom MATLAB code (The Mathworks, Inc.; Natick, MA) was used to rotate the STLs of the humerus and supraspinatus tendon and footprint relative to the scapula according to these mean glenohumeral joint positions. The center of the humeral head was assumed to remain centered in the glenoid for all glenohumeral positions.

Subacromial proximity was quantified as the minimum distance between the CA arch (both acromial and CA ligament undersurfaces) and the supraspinatus tendon or footprint. A custom MATLAB code was used to calculate 3D Euclidean distance between every possible combination of vertices on the surface of the CA arch and the supraspinatus tendon or footprint. From these distances, a single minimum distance was identified for each glenohumeral position for both the tendon and footprint. The level of refinement of the surface meshes allowed for the minimum distance to be calculated at a resolution of approximately 6 vertices/mm2. When the minimum distance was determined to be less than 0 mm, the 3D volume of the supraspinatus tendon that intersected with the CA arch was calculated using a Boolean intersection method within Mimics.

Because the same magnitude of humeral retroversion was imposed on each subjects’ 3D model, a sensitivity analysis was performed to investigate the impact of different magnitudes of retroversion on the minimum distance and volume of intersection. At each glenohumeral joint position, adjustments in retroversion were imposed by increasing and decreasing the retroversion angle by 20° and 30°. These values represent approximately ±1.5 and ±2.0 standard deviations from the mean retroversion angle (57.2° ± 14°).29, 30

Statistical Analysis

Descriptive statistics were calculated at 5° increments of humerothoracic elevation for each dependent variable (minimum distance and volume of intersection). All statistical analyses were performed using SAS 9.4 with an a priori Type I error rate of 0.05 (SAS Institute, Inc., Cary, NC). Demographics were compared using independent t-tests for continuous data and Fisher’s Exact test for proportions.

The assumption of normality was assessed prior to performing all statistical analyses by assessing skewness and kurtosis. The sample distributions for minimum distance to both the supraspinatus tendon and footprint were positively skewed in the asymptomatic group at each combination of humerothoracic elevation and retroversion adjustment. Transformation of the dataset was unsuccessful as it resulted in severe skewness in the symptomatic group. Upon closer inspection, the data for one asymptomatic subject was consistently more than 2 standard deviations from the mean of the asymptomatic group. Removing this subject resulted in skewness values within acceptable levels and subsequent parametric analyses were performed with the remaining nine asymptomatic subjects. The subject was retained for descriptive analyses and is discussed later. The sample distribution for the volume of the supraspinatus tendon that intersected with the CA arch was also highly skewed. Transformation of data was not successful. Therefore, the magnitude of volume intersections were ranked across humerothoracic elevation angles within each subject and a Friedman’s non-parametric ANOVA was performed to compare the mean ranks. This process was repeated for each magnitude of retroversion.

Three-factor mixed-model ANOVAs were performed to investigate the effects of humerothoracic elevation (4 levels: 0°, 30°, 60°, 90°), retroversion adjustment (5 levels: 0°, ±20°, ±30°), and group (2 levels: asymptomatic, symptomatic) on the minimum distance between the CA arch and the supraspinatus tendon and footprint. In the presence of a significant three-factor interaction, separate two-factor mixed-model ANOVAs (factors: group, humerothoracic elevation) were performed at each angle of retroversion. In the presence of a significant two-factor interaction, simple effects were examined at each level of the interacting factors using Bonferroni adjustments. When following up the effect of humerothoracic elevations, six comparisons were performed between the four humerothoracic elevation angles, resulting in a Bonferroni critical value of p=0.0083. When following-up the effect of retroversion adjustment, four comparisons were performed (0° vs. ±20° and ±30°) resulting in a Bonferroni critical value of p=0.0125.

It should be noted that group comparisons were not the primary aim of the study because equivalent kinematics were implemented across groups. However including group as a factor in the statistical analysis permitted potential anatomical differences between groups to be identified as a source of statistical variance in proximity measures.

Results

Demographics

No demographic differences were found between the symptomatic and asymptomatic groups (TABLE 1).

Minimum Distance to the Supraspinatus Tendon

When the mean humeral retroversion was imposed on the model (i.e. 0° retroversion adjustment), the minimum distance between the CA arch and the supraspinatus tendon generally followed a similar pattern across subjects (FIGURE 2). The minimum distance remained small at lower angles of humerothoracic elevation and reached the minima at an average of 42° humerothoracic elevation before gradually increasing until the end of the simulated motion.

FIGURE 2.

FIGURE 2

Open in a new tab

Individual subject plot for the minimum distance between the coracoacromial arch and the supraspinatus tendon across the range of motion when the mean humeral retroversion was imposed on the model (i.e. 0° retroversion adjustment).

The minimum distance to the supraspinatus tendon was significantly affected by both the humerothoracic elevation angle and humeral retroversion angle (two-factor interaction: p<0.0001, F=20.54, df=12,204) (FIGURE 3). When this interaction was followed-up to examine the effect of humerothoracic elevation at each retroversion magnitude, the smallest minimum distance occurred at 30° humerothoracic elevation (0.7–0.8 mm) across all retroversion magnitudes. However, the magnitude of the minimum distance at 30° humerothoracic elevation was not generally significantly different from that at 0° or 60° humerothoracic elevation.

FIGURE 3.

FIGURE 3

Open in a new tab

Effect of humerothoracic elevation angle on the minimum distance between the supraspinatus tendon and the coracoacromial arch across different magnitudes of humeral retroversion. Bars within the same retroversion angle that share the same letter are not signfiicantly different (p>0.0125).

When the interaction was followed-up to look at the effect of adjusting humeral retroversion at each angle of humerothoracic elevation, the minimum distance was found to be unaffected by increasing retroversion magnitudes (FIGURE 4). However, a decrease in humeral retroversion significantly increased the minimum distance at 60° and 90° humerothoracic elevation. Specifically, at 60° humerothoracic elevation, a 30° reduction in retroversion resulted in a 0.6 mm increase in the minimum distance (p=0.0005, t=3.56, df=204). At 90° humerothoracic elevation, a 20° reduction in retroversion resulted in a 0.9 mm increase in the minimum distance (p<0.0001, t=5.78, df=204), while a 30° reduction resulted in a 1.7 mm increase in the minimum distance (p<0.0001, t=10.65, df=204).

FIGURE 4.

FIGURE 4

Open in a new tab

Effect of different humeral retroversion magnitues on the minimum distance between the supraspinatus tendon and the coracoacromial arch across humerothoracic elevation. *A 30° reduction in retroversion resulted in a significantly larger minimum distance than the mean retroversion (0°). † Both a 30° and 20° reduction in retroversion resulted in significantly larger minimum distances than the mean retroversion (0°).

Minimum Distance to the Supraspinatus Footprint

When the mean humeral retroversion was imposed on the model (i.e. 0° retroversion adjustment), the minimum distance between the CA arch and the supraspinatus footprint generally followed a similar pattern across subjects (FIGURE 5). The minimum distance decreased with increasing angles of humeral elevation until a minima occurred at an average of 56° humerothoracic elevation. After this point, the minimum distance gradually increased until the end of the simulated motion.

FIGURE 5.

FIGURE 5

Open in a new tab

Individual subject plot for the minimum distance between the coracoacromial arch and the supraspinatus footprint across the range of motion when the mean humeral retroversion was imposed on the model (i.e. 0° retroversion adjustment).

The minimum distance to the supraspinatus footprint was significantly affected by the humerothoracic elevation angle, humeral retroversion magnitude, and group (three-factor interaction: p=0.0155, F=2.15, df=12,204). When the mean humeral retroversion was imposed on the model or decreased by 20°, the difference between humerothoracic elevation angles depended on group (two-factor interactions: p=0.0080, F=4.39, df=3,51 and p=0.0182, F=3.66, df=3,51, respectively) (FIGURES 6 and S-1). The average minimum distance was smallest at 60° humerothoracic elevation in both groups for both humeral retroversion angles. However these magnitudes were not significantly different from 30° or 90° humerothoracic elevation. Additionally, at 0° humerothoracic elevation, the symptomatic group had 2.9 mm greater minimum distance than the asymptomatic group (p=0.0027, t=−3.15, df=51) when the mean retroversion was imposed on the model, and 2.6 mm greater minimum distance than the asymptomatic group when humeral retroversion was decreased by 20° (p=0.0095, t=−2.69, df=51).

FIGURE 6.

FIGURE 6

Open in a new tab

Effect of humeral elevation angle on the minimum distance between the supraspinatus footprint and the CA arch when the mean retroversion was imposed on the model (0° retroversion adjustment). Bars within the same group that share the same letter are not signfiicantly different (p>0.0125).

When the retroversion magnitude was decreased by 30° or increased by 20° or 30°, the minimum distance was affected only by humerothoracic elevation angle (main effects: p<0.0001, F=12.94, df=3,51; p<0.0001, F=34.56, df=3,51; p<0.0001, F=27.07, df=3,51, respectively) (FIGURE S-2). At these retroversion magnitudes, the average minimum distance was smallest at 60° humerothoracic elevation. However, the magnitude of the minimum distance did generally not differ significantly from that at 30° or 90° humerothoracic elevation.

Volume Intersection between CA Arch and Supraspinatus Tendon

Descriptively, the largest volume of intersection throughout the simulated reaching task occurred at 30° humerothoracic elevation across all retroversion angles. However, no significant differences existed between angles of humerothoracic elevation (TABLE S-4). When the mean humeral retroversion was imposed on the humeral model (i.e. 0° retroversion adjustment), intersections between the supraspinatus tendon and CA arch occurred in 4/10 asymptomatic subject models and 6/10 symptomatic subject models when considered over the entire simulated range of motion (30 glenohumeral positions between 0°–150° humerothoracic elevation) (FIGURE S-3). In models where the supraspinatus tendon intersected with the CA arch during the simulated reaching task, intersections occurred 7 to 17 times for those in the asymptomatic group (median: 10) and 1 to 21 times for those in the symptomatic group (median: 7).

Discussion

The pathogenesis of rotator cuff disease is generally believed to be multi-factorial with theories ranging from intrinsic tendon degeneration and reduced vascularity to extrinsic tendon injury from anatomical and biomechanical factors.4, 5 Orthopaedic clinicians are often focused on addressing the extrinsic mechanisms through surgical and rehabilitative treatments. The results of the current study provide support for the extrinsic or mechanical theory of rotator cuff disease development by demonstrating the effect of glenohumeral elevation (kinematic mechanism) on measures of supraspinatus rotator cuff compression. Standardized kinematics were imposed on the anatomical models, and compression of the supraspinatus under the CA arch was observed even in the models of asymptomatic individuals. This finding is important to consider as a centered humeral head and average healthy subject kinematics likely represent “optimal” kinematics. Deviations from these values could further compromise the rotator cuff in these elevation positions at highest risk.

Further, the range of motion in which the supraspinatus footprint was at risk for compression (30°–90° humerothoracic elevation) was lower in the range of motion than previously reported (near or above 90° humerothoracic elevation).12, 13, 15, 31 However, these previous studies quantified proximity to the humeral head rather than directly to the rotator cuff. Consequently, they cannot account for the fact that at increasing angles of humerothoracic elevation, the acromiohumeral distance vector moves progressively more lateral, contacts the lateral humerus, and no longer accurately represents risk to the footprint (FIGURE S-4).15 This phenomenon was observed in nearly all subjects in the current study as the supraspinatus footprint cleared the acromion and began to move further from the CA arch by approximately 90° (FIGURE 5). This finding has important implications for how rotator cuff injury risk is theorized to occur relative to overhead work, and how higher range of motion “impingement” tests (e.g. Neer test) are interpreted.

Compared to the footprint, risk for compression of the supraspinatus tendon occurs at even lower angles, between 0° and 60° humerothoracic elevation. Mechanically, the shift in risk is likely because the tendon includes portions of the medial supraspinatus (FIGURE 1) that may approximate the CA arch before the footprint during humeral elevation. Clinically, this finding could be useful in identifying positions or arcs of motion that should be avoided to protect the supraspinatus from mechanical compression as rotator cuff tendon tears are often initiated in the musculotendinous junction.

Clinically, mechanical subacromial compression was presumed to occur at higher angles of elevation due to the prevalence of symptom provocation within this range. However, the results of this study suggest pain provocation is likely the result of other mechanisms. For example, tendon injury due to mechanical subacromial compression occurring at lower angles may result in symptoms at higher angles where the muscle force required increases to counteract the higher external torque. Additionally, symptoms at overhead positions of arm elevation may be the result of entrapment of the undersurface of the supraspinatus (i.e. internal or posterior impingement32) rather than subacromial compression.33, 34 Further studies are needed to explore the relationship between clinical symptom presentation, biomechanical factors, sources of rotator cuff mechanical compression, and other potential pain generators.

The use of subject-specific anatomic models allowed for investigation into potential soft tissue deformation when the supraspinatus tendon makes contact with the CA arch. However, because the tendons were rotated as rigid bodies in this analysis, deformation was not directly assessed. Therefore, the volume of supraspinatus tendon that intersected with the CA arch was quantified as a proxy because the degree to which the rigid bodies overlap should represent the magnitude of soft tissue that would have to deform as it comes into contact with the CA arch. While the statistical assessment of the volume of intersection was hindered because contact occurred in only 50% of the subjects, the relative pattern was similar to that of minimum distances. Future studies using finite element modeling could quantify rotator cuff deformation and would provide an optimal assessment of rotator cuff injury risk.

Minimum distances were also quantified using varying magnitudes of humeral retroversion to determine how sensitive the minimum distance measure was to the use of a single humeral retroversion angle across subjects. Interestingly, altering the magnitude of retroversion generally did not significantly alter subacromial proximities in the range of motion where the rotator cuff is at risk for compression. These findings may also shed light on the impact of humeral axial rotation on supraspinatus compression risk as an increase in retroversion is mechanically identical to an increase in glenohumeral internal rotation. Clinically, decreased glenohumeral external rotation (or relative internal rotation) is thought to contribute to subacromial impingement during arm elevation because the rotator cuff is not allowed to clear the acromion.35 However, the findings of the current study suggest decreased glenohumeral external rotation (i.e. increased retroversion) does not significantly affect the minimum distance between the supraspinatus and CA arch throughout the range of motion. This is likely because reducing glenohumeral external rotation only shifted where the minimum distance vector contacted the surface of the tendon. In contrast, externally rotating the humerus at higher angles of humeral elevation increased the minimum distance by rotating the supraspinatus tendon further away from the CA arch. Therefore, reducing glenohumeral external rotation (or relative internal rotation) may not impact subacromial clearance, but increasing external rotation may be protective at higher angles of arm elevation.

In addition to shoulder kinematics, variations in anatomy have been proposed as an extrinsic factor of mechanical subacromial rotator cuff compression.4, 5 In the current study, group differences in minimum distance would suggest anatomical variances exist between groups because the same glenohumeral kinematics were imposed on all of the anatomical models for a given humeral elevation position. Interestingly, the only significant group difference observed indicated the supraspinatus footprint in asymptomatic subject models was closer to the CA arch than for symptomatic subject models at 0° humerothoracic elevation. By definition, this position occurs when the long axis of the humerus is parallel to the vertical axis of the trunk. The shoulder is unlikely to be in this position even in a relaxed standing posture due to the width of the trunk. Therefore while statistically significant, there may be no clinical meaningfulness to this group difference. Furthermore, differences between groups in the presence of rotator cuff pathology were not reported because the MR images were acquired using parameters optimized for anatomical identification and not the presence of inflammation or pathology.

When considering individual subject plots (FIGURES 2 and 5), the results of the study also suggest between-subject variation in anatomy influences potential for rotator cuff compression during arm elevation, but in more complex ways than historically suggested. For example, acromial slope has been an anatomical variant most commonly theorized to increase risk for rotator cuff compression and injury.36 Exploration of the outlier in this study provides perspective for future studies seeking to explore how variations in shoulder joint anatomy impact subacromial rotator cuff compression. The asymptomatic subject that was excluded from the statistical analysis of minimum distances (FIGURES 2 and 5) was 70 years old and had no previous episode of shoulder pain despite playing recreational softball throughout her life. Inspection of the subject’s models suggests the subject’s anatomy may have been protective by allowing larger distance between the supraspinatus and CA arch than what was observed in other subjects (FIGURE S-5). Future investigations using individual subject kinematics and quantifying anatomical factors such as acromial slope, acromial lateralization, humeral head radius, glenoid inclination, and glenoid version may shed light on the likely complex relationships between shoulder kinematics and anatomy.

When interpreting the results of the study, it is important to consider how kinematics were imposed on the anatomical models. The use of standardized kinematics precludes the interpretation of the data in terms of what happens when an individual subject moves. However, imposing subject-specific kinematics on the models would have confounded the planned systematic investigation into the effect of humeral elevation on subacromial proximities due to the inherent between-subject variability in kinematics. Instead, average asymptomatic angular kinematics without translations were imposed on the models which can be considered a best estimate of “optimal” kinematics.

The average between-subject variability (i.e. SD) for both glenohumeral axial rotation and plane of elevation during the functional reaching task was approximately 10°.26 This between-subject variability in axial rotation was accounted for in the retroversion sensitivity analysis, which suggest retroversion variability (and thus typical between-subject axial rotation variability) would not significantly impact the minimum distance. An exploratory sensitivity analysis was also performed using the anatomical model of one subject to investigate the impact of the variability in glenohumeral plane of elevation. In this analysis, plane of elevation was altered by ±10° and ±20° (±1 and ±2 SD, respectively). These substantial variations in plane of elevation were found to only impact the minimum distance by an average magnitude (RMS) of 0.4 mm compared to the standardized kinematics used in this study. Considered together, these findings suggest the between-subject variability in glenohumeral axial rotation and plane of elevation expected for 95% of the population (i.e. 2 SD) does not significantly impact subacromial proximity measures at the humerothoracic angles assessed statistically. Perhaps more importantly, the analysis also suggests the use of standardized kinematics in this study did not significantly impact the descriptions of subacromial proximity.

This study has limitations that should be considered when interpreting the results. First, the segmentation process to create the 3D anatomical models largely relied on manual segmentation. While this was more time consuming than using an automated process, manually identifying pixels to include in a structure’s mask allowed for the attempt to identify the individual rotator cuff tendons as they tend to blend at their insertion. We estimate the typical segmentation error was 1 pixel, with a worst-case error of 2 pixels. Given the 0.6 mm isotropic resolution of the MR scan, it is therefore believed the accuracy of the models to be within 1 mm or less.

Second, like all modeling studies several assumptions were made to control for unknown parameters. A mean retroversion angle of 57.2°25 was used because the humeral epicondyles were not visible to construct the anatomical coordinate systems. However, constraining this variable allowed for the direct assessment of the impact of humeral elevation on proximity measures without the confounding effect of between-subject retroversion variability. Furthermore, the results of our sensitivity analysis indicated this assumption did not significantly influence the magnitude of the minimum distance to the supraspinatus tendon at or below 60° humerothoracic elevation for 87% of the population (±1.5 SD).

Third, the humeral head was assumed to remain centered in the glenoid to directly assess the impact of humeral elevation without the confounding effects of glenohumeral translation. Controlling glenohumeral translation and between-subject variability by imposing standardized kinematics was a necessary methodological step to avoid the influence of other confounding variables and address the study’s primary aim to examine the effect of altering specific kinematic factors (humeral elevation and retroversion). Modeling without translations also reflects “optimal” kinematics because only angular motion occurs at the joint. Previous studies show the glenohumeral joint may not function in this ideal way as the humerus translates slightly relative to the glenoid during arm motion.37, 38 As such, superior translation would likely decrease the minimum distance and increase the volume of intersection, while inferior translation would likely result in the inverse.

Lastly, the 3D anatomic models were constructed from MR images taken with the subject in a relaxed supine position. Muscle contraction may alter the thickness of the supraspinatus tendon and resulting proximity measures. However if the supraspinatus were contracted as expected throughout the motion, this alteration would likely be a fixed offset from the values reported.

Ultimately this study provides evidence for mechanical rotator cuff compression during low angles of functional arm elevation as a potential mechanism for the development of rotator cuff disease. Risk for supraspinatus compression was particularly greatest between 0° and 90° humerothoracic elevation, which is lower than previously thought. This information can be used by clinicians to help inform activity modification and exercise prescription to avoid shoulder positions that may increase the risk of supraspinatus compression. Future research is needed to advance the investigation of rotator cuff disease from proximity measures to descriptions of tissue strain, to investigate the relationship between shoulder motion and variations in shoulder anatomy, and to determine how specific shoulder movement abnormalities affect compression risk.

Supplementary Material

Supp Fig S1-5
Supplementary Material

Acknowledgments

This project was supported by Award Number R03-HD070871 from the Eunice Kennedy Shriver National Institute of Child Health and Human Development (Ludewig). The content is solely the responsibility of the authors and does not necessarily represent the official views of the Eunice Kennedy Shriver National Institute of Child Health and Human Development or the National Institutes of Health. This project was also supported by a pre-doctoral training grant (T32-AR050938) through the National Institute of Arthritis and Musculoskeletal and Skin Diseases (Lawrence), a pre-doctoral fellowship (F31-HD087069) through the National Institute of Child Health and Human Development (Lawrence), and pre-doctoral scholarships through the Foundation for Physical Therapy (Lawrence and Staker). One author (Braman) receives research funding from Stryker Corporation. No funds received from Stryker were used for this project. One author (Lawrence) received an NIH Loan Repayment Program award L30-HD089226 from the National Institutes of Child Health and Human Development.

Footnotes

Author Contributions Statement: All authors substantially contributed to the study and have read and approved the final submitted manuscript. Specific author contributions are as follows:
  • Research design: Lawrence, Schlangen, Schneider, Schoenecker, Senger, Starr, Ellermann, Braman, Ludewig
  • Data acquisition: all authors
  • Data analysis: Lawrence, Schlangen, Schneider, Schoenecker, Senger, Starr, Staker, Braman, Ludewig
  • Data interpretation: all authors
  • Drafting and revising the paper: all authors

References

  • 1.Picavet HSJ, Schouten JSAG. Musculoskeletal pain in the Netherlands: Prevalences, consequences and risk groups, the DMC(3)-study. Pain. 2003;102:167–178. doi: 10.1016/s0304-3959(02)00372-x. [DOI] [PubMed] [Google Scholar]
  • 2.Roe Y, Bautz Holter E, Juel N, Soberg H. Identification of relevant international classification of functioning, disability and health categories in patients with shoulder pain: A cross-sectional study. J Rehabil Med. 2013;45:662–9. doi: 10.2340/16501977-1159. [DOI] [PubMed] [Google Scholar]
  • 3.Freygant M, Dziurzyńska-Białek E, Guz W, et al. Magnetic resonance imaging of rotator cuff tears in shoulder impingement syndrome. Pol J Radiol. 2014;79:391–397. doi: 10.12659/PJR.890541. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Michener LA, McClure PW, Karduna AR. Anatomical and biomechanical mechanisms of subacromial impingement syndrome. Clin Biomech. 2003;18:369–379. doi: 10.1016/s0268-0033(03)00047-0. [DOI] [PubMed] [Google Scholar]
  • 5.Seitz AL, McClure PW, Finucane S, et al. Mechanisms of rotator cuff tendinopathy: Intrinsic, extrinsic, or both? Clin Biomech (Bristol, Avon) 2011;26:1–12. doi: 10.1016/j.clinbiomech.2010.08.001. [DOI] [PubMed] [Google Scholar]
  • 6.Neer CS. Impingement lesions. Clin Orthop. 1983:70–77. [PubMed] [Google Scholar]
  • 7.Waldt S, Bruegel M, Mueller D, et al. Rotator cuff tears: Assessment with MR arthrography in 275 patients with arthroscopic correlation. Eur Radiol. 2007;17:491–498. doi: 10.1007/s00330-006-0370-7. [DOI] [PubMed] [Google Scholar]
  • 8.Schellingerhout JM, Verhagen AP, Thomas S, Koes BW. Lack of uniformity in diagnostic labeling of shoulder pain: Time for a different approach. Man Ther. 2008;13:478–483. doi: 10.1016/j.math.2008.04.005. [DOI] [PubMed] [Google Scholar]
  • 9.Papadonikolakis A, McKenna M, Warme W, et al. Published evidence relevant to the diagnosis of impingement syndrome of the shoulder. J Bone Joint Surg Am. 2011;93:1827–1832. doi: 10.2106/JBJS.J.01748. [DOI] [PubMed] [Google Scholar]
  • 10.Kibler WB, Ludewig P, McClure P, et al. Clinical implications of scapular dyskinesis in shoulder injury: The 2013 consensus statement from the ‘scapular summit’. Br J Sports Med. 2013;47:877–885. doi: 10.1136/bjsports-2013-092425. [DOI] [PubMed] [Google Scholar]
  • 11.Braman JP, Zhao KD, Lawrence RL, et al. Shoulder impingement revisited: Evolution of diagnostic understanding in orthopaedic surgery and physical therapy. Med Biol Eng Comput. 2014;52:211–219. doi: 10.1007/s11517-013-1074-1. [DOI] [PubMed] [Google Scholar]
  • 12.Graichen H, Bonel H, Stammberger T, et al. Three-dimensional analysis of the width of the subacromial space in healthy subjects and patients with impingement syndrome. AJR Am J Roentgenol. 1999;172:1081–1086. doi: 10.2214/ajr.172.4.10587151. [DOI] [PubMed] [Google Scholar]
  • 13.Bey M, Brock S, Beierwaltes W, et al. In vivo measurement of subacromial space width during shoulder elevation: Technique and preliminary results in patients following unilateral rotator cuff repair. Clin Biomech. 2007;22:767–73. doi: 10.1016/j.clinbiomech.2007.04.006. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Thompson MD, Landin D, Page PA. Dynamic acromiohumeral interval changes in baseball players during scaption exercises. J Shoulder Elbow Surg. 2011;20:251–258. doi: 10.1016/j.jse.2010.07.012. [DOI] [PubMed] [Google Scholar]
  • 15.Giphart JE, van der Meijden OA, Millett PJ. The effects of arm elevation on the 3-dimensional acromiohumeral distance: A biplane fluoroscopy study with normative data. J Shoulder Elbow Surg. 2012;21:1593–1600. doi: 10.1016/j.jse.2011.11.023. [DOI] [PubMed] [Google Scholar]
  • 16.Silva RT, Hartmann LG, Laurino CF, Biló JP, Rocha Clinical and ultrasonographic correlation between scapular dyskinesia and subacromial space measurement among junior elite tennis players. Br J Sports Med. 2010;44:407–410. doi: 10.1136/bjsm.2008.046284. [DOI] [PubMed] [Google Scholar]
  • 17.Seitz A, McClure P, Lynch S, et al. Effects of scapular dyskinesis and scapular assistance test on subacromial space during static arm elevation. J Shoulder Elbow Surg. 2012;21:631–40. doi: 10.1016/j.jse.2011.01.008. [DOI] [PubMed] [Google Scholar]
  • 18.Hébert LJ, Moffet H, Dufour M, Moisan C. Acromiohumeral distance in a seated position in persons with impingement syndrome. Journal of magnetic resonance imaging. 2003;18:72–79. doi: 10.1002/jmri.10327. [DOI] [PubMed] [Google Scholar]
  • 19.Meyer AW. Chronic functional lesions of the shoulder. Arch Surg. 1937;35:646–674. [Google Scholar]
  • 20.Michener L, Walsworth M, Doukas W, Murphy K. Reliability and diagnostic accuracy of 5 physical examination tests and combination of tests for subacromial impingement. Arch Phys Med Rehabil. 2009;90:1898–1903. doi: 10.1016/j.apmr.2009.05.015. [DOI] [PubMed] [Google Scholar]
  • 21.Curtis A, Burbank K, Tierney J, et al. The insertional footprint of the rotator cuff: An anatomic study. Arthroscopy. 2006;22:609.e1–609.e1. doi: 10.1016/j.arthro.2006.04.001. [DOI] [PubMed] [Google Scholar]
  • 22.Jones AO. Magnetic resonance imaging of the supraspinatus tendon: The significance of signal intensity alterations at the ‘critical zone’. Australas radiol. 1998;42:106–113. doi: 10.1111/j.1440-1673.1998.tb00584.x. [DOI] [PubMed] [Google Scholar]
  • 23.Wu G, van der Helm FC, Veeger HE, et al. ISB recommendation on definitions of joint coordinate systems of various joints for the reporting of human joint motion--part II: Shoulder, elbow, wrist and hand. J Biomech. 2005;38:981–992. doi: 10.1016/j.jbiomech.2004.05.042. [DOI] [PubMed] [Google Scholar]
  • 24.Verstraeten TRGM, Deschepper E, Jacxsens M, et al. Determination of a reference system for the three-dimensional study of the glenohumeral relationship. Skeletal Radiol. 2013;42:1061–1071. doi: 10.1007/s00256-013-1572-0. [DOI] [PubMed] [Google Scholar]
  • 25.Massimini DF, Warner JJP, Li G. Glenohumeral joint cartilage contact in the healthy adult during scapular plane elevation depression with external humeral rotation. J Biomech. 2014;47:3100–3106. doi: 10.1016/j.jbiomech.2014.06.034. [DOI] [PubMed] [Google Scholar]
  • 26.Braman JP, Engel SC, Laprade RF, Ludewig PM. In vivo assessment of scapulohumeral rhythm during unconstrained overhead reaching in asymptomatic subjects. J Shoulder Elbow Surg. 2009;18:960–967. doi: 10.1016/j.jse.2009.02.001. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Ludewig PM, Hassett D, Laprade R, et al. Comparison of scapular local coordinate systems. Clin Biomech. 2010;25:415–421. doi: 10.1016/j.clinbiomech.2010.01.015. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Phadke V, Braman JP, LaPrade RF, Ludewig PM. Comparison of glenohumeral motion using different rotation sequences. J Biomech. 2011;44:700–705. doi: 10.1016/j.jbiomech.2010.10.042. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Kummer FJ, Perkins R, Zuckerman JD. The use of the bicipital groove for alignment of the humeral stem in shoulder arthroplasty. J Shoulder Elbow Surg. 1998;7:144–146. doi: 10.1016/s1058-2746(98)90225-7. [DOI] [PubMed] [Google Scholar]
  • 30.Balg F, Boulianne M, Boileau P. Bicipital groove orientation: Considerations for the retroversion of a prosthesis in fractures of the proximal humerus. J Shoulder Elbow Surg. 2006;15:195–198. doi: 10.1016/j.jse.2005.08.014. [DOI] [PubMed] [Google Scholar]
  • 31.Flatow EL, Soslowsky LJ, Ticker JB, et al. Excursion of the rotator cuff under the acromion: Patterns of subacromial contact. Am J Sports Med. 1994;22:779–788. doi: 10.1177/036354659402200609. [DOI] [PubMed] [Google Scholar]
  • 32.Walch G, Boileau P, Noel E, Donell ST. Impingement of the deep surface of the supraspinatus tendon on the posterosuperior glenoid rim: An arthroscopic study. J Shoulder Elbow Surg. 1992;1:238–245. doi: 10.1016/S1058-2746(09)80065-7. [DOI] [PubMed] [Google Scholar]
  • 33.Pappas GP, Blemker SS, Beaulieu CF, et al. In vivo anatomy of the neer and hawkins sign positions for shoulder impingement. J Shoulder Elbow Surg. 2006;15:40–49. doi: 10.1016/j.jse.2005.04.007. [DOI] [PubMed] [Google Scholar]
  • 34.Lawrence RL, Staker JL, Braman JP, Ludewig PM. Mechanical internal impingement of the supraspinatus tendon during a simulated reaching task. Proceedings of the 11th Conference of the International Shoulder Group; 2016. Available at: https://www.zhaw.ch/storage/engineering/institute-zentren/imes/schulterkongress_2016/Programme_and_Book_of_Abstracts_ISG_2016.pdf. [Google Scholar]
  • 35.Browne AO, Hoffmeyer P, Tanaka S, et al. Glenohumeral elevation studied in three dimensions. J Bone Joint Surg Br. 1990;72:843–845. doi: 10.1302/0301-620X.72B5.2211768. [DOI] [PubMed] [Google Scholar]
  • 36.Bigliani LU, Ticker JB, Flatow EL, Soslowsky LJ, Mow VC. The relationship of acromial architecture to rotator cuff disease. Clinics in sports medicine. 1991;10:823–838. [PubMed] [Google Scholar]
  • 37.Bey MJ, Kline SK, Zauel R, et al. Measuring dynamic in-vivo glenohumeral joint kinematics: Technique and preliminary results. J Biomech. 2008;41:711–714. doi: 10.1016/j.jbiomech.2007.09.029. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38.Ludewig PM, Cook TM. Translations of the humerus in persons with shoulder impingement symptoms. J Orthop Sports Phys Ther. 2002;32:248–259. doi: 10.2519/jospt.2002.32.6.248. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supp Fig S1-5
NIHMS843465-supplement-Supp_Fig_S1-5.pdf (487.9KB, pdf)
Supplementary Material
NIHMS843465-supplement-Supplementary_Material.doc (73KB, doc)

RESOURCES