The Coupled Kinematics of Scapulothoracic Upward Rotation

Abstract

Background

Scapulothoracic upward rotation (UR) is an important shoulder complex motion allowing for a larger functional work space and improved glenohumeral muscle function. However, the kinematic mechanisms producing scapulothoracic UR remain unclear, limiting the understanding of normal and abnormal shoulder movements.

Objective

The objective of this study was to identify the coupling relationships through which sternoclavicular and acromioclavicular joint motions contribute to scapulothoracic UR.

Design

This was a cross-sectional observational study.

Methods

Sixty participants were enrolled in this study; 30 had current shoulder pain, and 30 had no history of shoulder symptoms. Shoulder complex kinematics were quantified using single-plane fluoroscopy and 2D/3D shape matching and were described as finite helical displacements for 30-degree phases of humerothoracic elevation (30 degrees–60 degrees, 60 degrees–90 degrees, and 90 degrees–120 degrees). A coupling function was derived to estimate scapulothoracic UR from its component motions of acromioclavicular UR, sternoclavicular posterior rotation, and sternoclavicular elevation as a function of acromioclavicular internal rotation. The proportional contributions of each of the component motions were also calculated and compared between phases of humerothoracic elevation and groups.

Results

Scapulothoracic UR displacement could be effectively predicted using the derived coupling function. During the 30- to 60-degree humerothoracic elevation phase, acromioclavicular UR accounted for 84.2% of scapulothoracic UR, whereas sternoclavicular posterior rotation and elevation each accounted for < 10%. During later phases, acromioclavicular UR and sternoclavicular posterior rotation each accounted for 32% to 42%, whereas sternoclavicular elevation accounted for < 11%.

Limitations

Error due to the tracking of sternoclavicular posterior rotation may have resulted in an underprediction of its proportional contribution and an overprediction of the proportional contribution of acromioclavicular UR.

Conclusions

Acromioclavicular UR and sternoclavicular posterior rotation are the predominant component motions of scapulothoracic UR. More research is needed to investigate how these coupling relationships are affected by muscle function and influenced by scapular dyskinesis.

Optimal shoulder complex function depends on a combination of glenohumeral, sternoclavicular, acromioclavicular, and scapulothoracic motion. In particular, scapulothoracic upward rotation (UR)—sometimes termed “lateral rotation”^{1  ,  2}—plays several important roles in shoulder function, including orienting the glenoid to promote glenohumeral joint congruency and maintaining glenohumeral muscle length to maximize contractile function.^{1  ,  3} Scapulothoracic UR also increases overall shoulder complex range of motion during arm raising via scapulohumeral rhythm.^{4  ,  5} Given their importance, scapulothoracic UR and glenohumeral elevation are among the most studied shoulder complex motions.^4–13

Clinically, altered scapulothoracic UR during arm raising has been found in individuals with adhesive capsulitis,^14–16 rotator cuff tears,^{17  ,  18} glenohumeral instability,^{19  ,  20} and osteoarthritis.^{16  ,  21} When motions deemed “abnormal” are observed during a clinical examination, exercises are often prescribed to address underlying impairments such as muscle function. However, selecting appropriate exercises for abnormal scapulothoracic motion can be challenging because the kinematic chain (ie, mechanical linkage) of the shoulder complex requires “coupled” scapular and clavicular motion. In other words, scapulothoracic motion occurs due to sternoclavicular and acromioclavicular joint motions occurring together.^{4  ,  22–25} As such, scapulothoracic muscles function through the sternoclavicular and acromioclavicular joints. Although assessing sternoclavicular and acromioclavicular motions is challenging with current clinical tools (eg, inclinometer), diagnosing and treating shoulder motion abnormalities is ultimately dependent upon a strong understanding of shoulder complex coupling.

Scapulothoracic motion results from sternoclavicular and acromioclavicular coupling that occurs simultaneously in all 3 dimensions. However, not all sternoclavicular and acromioclavicular joint motions contribute to a specific (or uniaxial) scapulothoracic motion (eg, UR). For example, of the 6 total acromioclavicular and sternoclavicular joint rotations, it has been proposed that only acromioclavicular UR, sternoclavicular posterior rotation, and sternoclavicular elevation contribute to scapulothoracic UR (Fig. 1).²⁴ Therefore, these 3 rotations are collectively termed the “component motions” of scapulothoracic UR.²⁴ However, the manner in which these component motions contribute to scapulothoracic UR is not additive. Instead, Teece et al²⁴ proposed that the degree to which these component motions contribute to scapulothoracic UR is a function of the offset between the scapular and clavicular medial-lateral axes (ie, the acromioclavicular internal rotation angle).⁵ However, this theory has not been directly tested. Thus, it remains unclear the extent to which the component motions contribute to scapulothoracic UR, resulting in gaps in our understanding of how to identify and treat shoulder pain associated with abnormal scapular movement.

Figure 1.

Definitions of shoulder complex motions. Scapulothoracic upward rotation: scapular rotation relative to the thorax about an axis perpendicular to the plane of the scapula, such that the glenoid faces more superiorly. Acromioclavicular upward rotation: scapular rotation relative to the clavicle about an axis perpendicular to the plane of the scapula, such that the glenoid faces more superiorly. Sternoclavicular posterior rotation: clavicular rotation relative to the thorax about the long axis of the clavicle, such that the superior clavicular surface faces more posteriorly. Sternoclavicular elevation: clavicular rotation relative to the thorax about an approximately anterior-posterior axis perpendicular to the long axis of the clavicle, such that the lateral clavicle moves superiorly. Collectively, acromioclavicular upward rotation, sternoclavicular posterior rotation, and sternoclavicular elevation are termed “component motions” of scapulothoracic upward rotation because they couple (ie, occur together) to produce scapulothoracic upward rotation. Humerothoracic elevation: elevation of the humerus relative to the thorax (red line). Scapulothoracic external rotation: scapular rotation relative to the thorax about the scapular superior-inferior axis, such that the glenoid faces more laterally. Scapulothoracic posterior tilt: scapular rotation relative to the thorax about the scapular medial-lateral axis, such that the posterolateral acromion moves posteriorly and the inferior angle moves anteriorly.

A major barrier to understanding shoulder complex coupling has been the challenges associated with quantifying clavicular kinematics. Despite this challenge, previous work provided construct validity to the coupling theory by showing group mean differences in scapulothoracic UR between symptomatic and asymptomatic participants could be partially explained by concurrent mean differences in component motion kinematics.⁷ However, the small sample size (due to the use of bone-fixed motion sensors) and high between-subject variability in movement patterns limits the utility of using group means to rigorously test the coupling theory. Recently, single-plane fluoroscopy and 2D/3D shape matching were shown to be valid methods for quantifying shoulder complex kinematics²⁶ and are noninvasive, allowing for the testing of a larger sample. Therefore, this methodology may help overcome previous methodological barriers to studying shoulder complex coupling.

The purpose of this study was to identify the coupling relationships through which sternoclavicular and acromioclavicular joint motions contribute to scapulothoracic UR during scapular plane abduction by investigating the theory proposed by Ludewig et al.²⁴ It was hypothesized that acromioclavicular UR would be the predominant component motion of scapulothoracic UR.

Methods

Participants

Sixty participants were enrolled in this exploratory study; 30 had current shoulder symptoms consistent with a clinical diagnosis of impingement syndrome,²⁷ and 30 had no history of shoulder pain. Individuals with and without shoulder pain were included to ensure a broad distribution of shoulder complex kinematics. Individuals were excluded if they were unable to achieve 120 degrees of humerothoracic elevation (ie, humeral elevation relative to the thorax) or had a history of rotator cuff tear, adhesive capsulitis, glenohumeral dislocation, acromioclavicular separation, or shoulder fracture or surgery. Symptomatic and asymptomatic groups were matched for age, sex, and the dominance of the side tested. Demographic data are presented in the Table. The study was approved by the Institutional Review Board and All-University Radiation Protection Advisory Committee at the University of Minnesota. All participants provided written informed consent.

Table.

Participant Demographics^a

Characteristic	Asymptomatic Group (n = 30)	Symptomatic Group (n = 30)	P
Age, y (mean ± SD)	32.7 ± 8.3	32.4 ± 8.8	.90
Sex (% men)	46.7	46.7	1.0
Dominance of the side tested (% dominant)	40.0	40.0	1.0
Height, cm (mean ± SD)	173.1 ± 8.5	170.9 ± 7.8	.32
Mass, kg (mean ± SD)	72.1 ± 12.5	72.9 ± 15.6	.82
BMI, kg/m2 (mean ± SD)	23.9 ± 2.7	24.8 ± 4.1	.33
Symptom duration, wk, median (IQR)	0.0 (0–0)	36 (18–104)	NC
NPRS score (0–100)
Highest in past 7 d (mean ± SD)	0 ± 0	49.6 ± 16.9	NC
Lowest in past 7 d, median (IQR)	0 (0–0)	0 (10–20)	NC
DASH score (0–100) (mean ± SD)	0.7 ± 1.8	15.3 ± 8.1	<.01
Work subscale, median (IQR)	0 (0–0)	0 (0–18.8)	NC
Sport subscale, median (IQR)	0 (0–0)	18.8 (3.2–37.5)	NC

^a Group comparisons were assessed using 2-sample independent t tests, Mann-Whitney U tests, or chi-square tests, as appropriate. BMI = body mass index; DASH = Disabilities of the Arm, Shoulder, and Hand; IQR = interquartile range (25th–75th); NC = not computed because of no variability or range in the asymptomatic group; NPRS = numeric pain rating scale.

Data Collection

Shoulder complex kinematics were collected while participants performed active scapular plane abduction (ie, 40° anterior to the coronal plane^{5  ,  28}) by spatially and temporally syncing a single-plane fluoroscopy system (Phillips BV Pulsera, Amsterdam, the Netherlands) and a 5-camera motion capture system (Vicon Motion Capture Systems, Hauppauge, NY, USA) using MotionMonitor software (Innovative Sports Training, Inc, Chicago, IL, USA) as described elsewhere.²⁹ In brief, participants were seated with the scapula approximately parallel to the radiographic image intensifier. Two rigid-body marker clusters, each consisting of 4 reflective markers, were secured over the sternum and the distal humerus. The arbitrary reference frames of both marker clusters were transformed to clinically meaningful coordinate systems using recommended standards.³⁰ During the motion trials, scapular and clavicular kinematics were tracked using the fluoroscopy system, while trunk and humeral kinematics were tracked using the camera system. Finally, shoulder magnetic resonance scans of the glenohumeral (ie, scapula and humerus) and clavicular regions were acquired using a Siemens MAGNETOM Prisma 3 T scanner (Siemens Healthcare, Erlangen, Germany) as previously described.^{29  ,  31}

Data Processing

Three-dimensional models of the scapula and clavicle were reconstructed from the magnetic resonance scans using Mimics software (Materialise NV, Leuven, Belgium). The scapular coordinate system was defined by digitizing anatomical landmarks.³² The clavicular coordinate system was defined by digitizing the midpoints of the proximal and distal ends and an arbitrary third point superior to the clavicle. The clavicular coordinate system was reoriented during post-processing by aligning its superior-inferior axis to that of the thorax when the participant’s arm was resting at his or her side.^{5  ,  7  ,  30}

XMALab software version 1.3.3³³ was used for fluoroscopic image calibration and distortion correction. Fluoroscopic images were down-sampled to every 10 degrees of humerothoracic elevation. JointTrack software³⁴ was used to track shoulder complex kinematics through a process called 2D/3D shape matching. Scapular and clavicular angular root-mean-square (RMS) errors associated with the protocol of our laboratory are 0.5 to 1.6 degrees and 0.4 to 3.7 degrees, respectively.²⁶

Sternoclavicular (clavicle relative to the thorax), acromioclavicular (scapula relative to the clavicle), and scapulothoracic (scapula relative to the thorax) kinematics were described as joint angular positions (Y-X′-Z′′) and finite helical displacements of the distal segment moving relative to the proximal segment for each 30-degree phase of humerothoracic elevation (ie, 30 degrees–60 degrees, 60 degrees–90 degrees, and 90 degrees–120 degrees).³⁵ The total helical rotational displacement represents the change in joint orientation about an oblique 3-dimensional axis between finite moments in time.³⁶ This total helical rotation was subsequently parsed into 3 helical angles relative to the coordinate system of the distal segment at the initial position of the phase.

Coupling Function Derivation

Scapulothoracic motion cannot be considered a simple summation of sternoclavicular and acromioclavicular joint motion.^{24  ,  37} Acromioclavicular motion will directly transfer to scapulothoracic motion because both describe scapular motion despite different proximal reference frames (ie, clavicle for acromioclavicular motion, thorax for scapulothoracic motion). However, sternoclavicular motion will not directly transfer to scapulothoracic motion because the axes of the clavicle and scapula are oblique to one another. Consequently, a uniaxial sternoclavicular motion will result in multiaxial scapulothoracic motion, preventing shoulder complex coupling from being simply additive.

Instead, Ludewig et al proposed that the sternoclavicular contribution to net scapulothoracic motion (via coupling) is a function of the acromioclavicular internal rotation angle, which represents the primary source of obliquity between the scapular and clavicular axes when the arm is at the side.²⁴ This physiological configuration results in a complex relationship between sternoclavicular and scapulothoracic motion to produce scapulothoracic UR. However, the researchers proposed the coupling relationships that result from the physiological configuration can be conceptually simplified using 2 “thought experiments” in which the acromioclavicular joint is oriented in 2 theoretical configurations (Fig. 2). Collectively, these configurations define the theoretical limits of the sternoclavicular contributions to scapulothoracic motion. For scapulothoracic UR, the 2 thought experiments are as follows.²⁴ First, if the acromioclavicular internal rotation angle was 0 degrees (ie, clavicular and scapular axes parallel), then the only sternoclavicular motion to produce scapulothoracic UR would be sternoclavicular elevation (Suppl. Video 1, available at https://academic.oup.com/ptj). Second, if the acromioclavicular internal rotation angle was 90 degrees (ie, clavicular and scapular axes perpendicular), then the only sternoclavicular motion to produce scapulothoracic UR would be sternoclavicular posterior rotation (Suppl. Video 2, available at https://academic.oup.com/ptj).

Figure 2.

Theoretical and average physiological configurations of the acromioclavicular joint axis defining the theoretical limits of how specific sternoclavicular motions contribute to scapulothoracic motion. The red line represents the clavicle medial-lateral axis, and the blue line represents the scapular medial-lateral axis. (A) Theoretical configuration in which the axes are parallel; therefore, scapulothoracic upward rotation is produced only by sternoclavicular elevation. (B) Theoretical configuration in which the axes are perpendicular; therefore, scapulothoracic upward rotation is produced only by sternoclavicular posterior rotation. (C) Physiological configuration with an acromioclavicular internal rotation angle of ~ 60°; scapulothoracic upward rotation is produced by both sternoclavicular posterior rotation and elevation. In addition to sternoclavicular motions, acromioclavicular upward rotation also contributes to scapulothoracic upward rotation.

In reality, however, the average physiological acromioclavicular internal rotation angle is approximately 60 degrees when the arm is resting at the side.⁵ Therefore, both sternoclavicular posterior rotation and elevation contribute to scapulothoracic UR (Suppl. Video 3, available at https://academic.oup.com/ptj) and the magnitudes of their contributions are—in part—a function of the acromioclavicular internal rotation angle.²⁴ More specifically, as the acromioclavicular internal rotation angle increases (ie, becomes more like the theoretical perpendicular axis configuration), the posterior rotation contribution increases while the elevation contribution decreases (and vice versa as the acromioclavicular internal rotation angle decreases). For example, for an acromioclavicular internal rotation angle of 60 degrees,⁵ the perpendicular axis configuration should account for approximately 67% of the physiological relationship between clavicular motion and scapular motion (ie, 60°/90° = 67%), with sternoclavicular posterior rotation serving as the primary sternoclavicular contributor to scapulothoracic UR. The parallel axis configuration should account for the remaining 33%, with sternoclavicular elevation further contributing to scapulothoracic UR.

To test the construct validity of the coupling theory, a coupling function was developed to predict each participant’s scapulothoracic UR from the component motions for each humerothoracic elevation phase (30°–60°, 60°–90°, and 90°–120°):

.In this equation, represents the predicted scapulothoracic UR helical displacement and , , and are the actual sternoclavicular elevation, sternoclavicular posterior rotation, and acromioclavicular UR helical displacements, respectively. The individual weighting factors for sternoclavicular elevation and sternoclavicular posterior rotation are a function of the mean acromioclavicular internal rotation angle () during each humerothoracic elevation phase, as follows:

and

The equation defines the extent to which the acromioclavicular internal rotation angle is coincident with the parallel axis configuration (ie, 0°) and thus estimates how much sternoclavicular elevation contributes to scapulothoracic UR (Suppl. Video 4, available at https://academic.oup.com/ptj).

The equation defines the extent to which the acromioclavicular internal rotation angle is coincident with the perpendicular axis configuration (ie, 90°) and thus estimates how much sternoclavicular posterior rotation contributes to scapulothoracic UR (Suppl. Video 4).

The weighting factor for acromioclavicular UR is 100% as scapular motion relative to the clavicle (ie, acromioclavicular) directly transfers to scapular motion relative to the thorax (ie, scapulothoracic) and is therefore independent of the acromioclavicular internal rotation angle.

From these weighting factors, it can be seen that if = 60 degrees (ie, average physiological configuration), then = 33% and = 67%, which are the expected results, as previously described (Suppl. Video 4).

Data Analysis

The accuracy of the coupling function for predicting the actual scapulothoracic UR displacement for each humerothoracic elevation phase was determined by first calculating errors for individual participants (predicted − actual). These errors were then aggregated across participants into RMS and bias (ie, mean) errors. The relationship between the actual and predicted scapulothoracic UR displacement was determined using r  ² from regression.

To explore the proportional (ie, relative) contributions of the component motions to scapulothoracic UR, the product of each component motion displacement and its weighting factor was expressed as a proportion of the actual scapulothoracic UR displacement, as follows:

 

and

Additionally, the residual error for prediction was expressed as a proportion of the actual scapulothoracic UR displacement. When summed for each participant, the 3 proportional contributions and residual error totaled 100%. Negative proportional contributions are possible, indicating the participant’s motion detracted from scapulothoracic UR (ie, sternoclavicular anterior rotation, sternoclavicular depression, acromioclavicular downward rotation).

The effects of humerothoracic elevation phase (30°–60°, 60°–90°, and 90°–120°) and group (symptomatic, asymptomatic) on component motion proportional contributions were tested using 2-factor mixed-model analyses of variance. The significance of the 2-factor interaction was assessed first. Significant main effects were only interpreted in the absence of a significant interaction. All follow-up comparisons were performed using Tukey adjustments. Statistical analyses were performed using SAS 9.4 (SAS Institute Inc, Cary, NC, USA) with alpha = .05.

Role of the Funding Source

Research reported in this article was supported by the Eunice Kennedy Shriver National Institute of Child Health and Human Development (F31-HD087069 and L30-HD089226), the National Institute of Arthritis and Musculoskeletal and Skin Diseases (T32-AR050938), the National Center for Advancing Translational Sciences (UL1-TR002494), and the National Institute of Biomedical Imaging and Bioengineering (P41-EB015894). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health (NIH). This work was also supported in part by Promotion of Doctoral Studies (PODS) Scholarships from the Foundation for Physical Therapy, the Minnesota Partnership for Biotechnology and Medical Genomics, a research infrastructure grant from the University of Minnesota Office of the Vice President for Research, and the University of Minnesota Department of Orthopaedic Surgery and Clinical and Translational Science Institute. Neither the NIH nor any of the other funders played a role in the design, conduct, or reporting of this study.

Results

On average across the range of motion, participants underwent scapulothoracic UR, acromioclavicular UR, sternoclavicular posterior rotation, sternoclavicular elevation, and a small amount of acromioclavicular internal rotation (Fig. 3).

Figure 3.

Joint orientations for acromioclavicular (AC) upward rotation, sternoclavicular (SC) posterior rotation, sternoclavicular elevation, scapulothoracic (ST) upward rotation, and acromioclavicular internal rotation during scapular plane abduction. Data are presented as the mean and unpooled SE.

Coupling Functions Predicting Scapulothoracic Upward Rotation

Across all humerothoracic elevation phases, the coupling function resulted in RMS errors of 1.2 to 2.8 degrees and bias errors of −2.4 to −0.5 degrees (Fig. 4). Further, the magnitude of scapulothoracic UR displacement predicted by the coupling function explained 80% to 89% of the variance in the actual scapulothoracic UR displacement (P < .01).

Figure 4.

Results of the coupling function for predicting scapulothoracic (ST) upward rotation (UR) displacement from the component motions of acromioclavicular upward rotation, sternoclavicular posterior rotation, and sternoclavicular elevation across humerothoracic elevation phases (30°–60°, 60°–90°, and 90°–120°). RMSE = root-mean-square error.

Proportional Contributions of Component Motions

Across all phases of humerothoracic elevation, acromioclavicular UR contributed the most to scapulothoracic UR (Fig. 5). The magnitude of contribution differed between humerothoracic elevation phases (main effect: P < .01; F  _2,115 = 33.30) without a significant interaction. During the 30- to 60-degree phase, acromioclavicular UR was responsible for an average of 84.2% of scapulothoracic UR. The proportional contribution decreased significantly to 42.3% for the 60- to 90-degree phase (P < .01; t  ₁₁₅ = 6.58), where it remained unchanged (36.6%) for the 90- to 120-degree phase (P = .64; t  ₁₁₅ = 0.90).

Figure 5.

Proportional contributions of sternoclavicular (SC) posterior rotation, sternoclavicular elevation, and acromioclavicular (AC) upward rotation (UR) across humerothoracic elevation phases (30°–60°, 60°–90°, and 90°–120°). The mean residual error in prediction is also presented as a proportion of scapulothoracic (ST) upward rotation.

Sternoclavicular posterior rotation was the secondary contributor to scapulothoracic UR (Fig. 5) and the magnitude of contribution differed between humerothoracic elevation phases (main effect: P < .01; F  _2,115 = 25.95) without a significant interaction. During the 30- to 60-degree phase, sternoclavicular posterior rotation was responsible for an average of 2.8% of scapulothoracic UR. The proportional contribution increased significantly to 32.2% for the 60- to 90-degree phase (P < .01; t  ₁₁₅ = −6.04), where it remained unchanged (34.2%) for the 90- to 120-degree phase (P = .91; t  ₁₁₅ = −0.40).

Sternoclavicular elevation consistently contributed little to scapulothoracic UR (Fig. 5). However, the magnitude of contribution differed between humerothoracic elevation phases (main effect: P = .02; F  _2,115 = 3.97) without a significant interaction. During the 30- to 60-degree and 60- to 90-degree phases, sternoclavicular elevation was responsible for an average of 8.3% and 8.6% of scapulothoracic UR, respectively (P = .93; t  ₁₁₅ = −0.37). During the 90- to 120-degree phase, the contribution increased significantly to 10.7% (P = .03; t  ₁₁₅ = −2.63).

Discussion

The results of this study support the theory proposed by Ludewig et al²⁴ that the coupled kinematics of scapulothoracic UR can be effectively described by 3 component motions (acromioclavicular UR, sternoclavicular posterior rotation, and sternoclavicular elevation) as a function of the acromioclavicular internal rotation angle. Although the derived coupling function only accounts for the transverse plane offset between the clavicular and scapular axes, this simplification appears to be reasonable at lower humerothoracic elevation angles. At higher angles, however, the scapular and clavicular axes become increasingly oblique to one another because of acromioclavicular UR and posterior tilt, necessitating a more complex coupling function to fully explain scapulothoracic UR. However, the average RMS error for predicting UR was only 2.8 degrees in the final phase. Further, the coupling function accounted for > 80% of component motion proportional contributions and explained a substantial portion of the variance (r  ² = 80%–89%) in actual scapulothoracic UR. Taken together, these findings suggest that the coupling function can simplify the complex motion of scapulothoracic UR into its component motions. Ultimately, a better understanding of coupling relationships may help inform the development of biomechanical theories to explain normal and abnormal scapulothoracic motions.

Given scapulothoracic UR can be effectively predicted from its component motions, the proportional contributions of the component motions can be explored. During the initial phase of humerothoracic elevation (ie, 30 degrees–60 degrees), acromioclavicular UR was the predominant component motion, accounting for an average of 84% of scapulothoracic UR. Sternoclavicular posterior rotation and elevation can be considered accessory motions during this phase with a combined average contribution of only 10%. The relative dominance of acromioclavicular UR and small contribution from sternoclavicular posterior rotation during this initial phase suggests that the scapula and clavicle are not yet fully coupled. In other words, scapular motion is not substantially transferred to the clavicle, likely because slack in the acromioclavicular joint and coracoclavicular ligaments needs to be reduced before motion can be transferred between bone structures.^{2  ,  38} Functionally, 1 potential benefit of not yet being fully coupled in the initial phase is that it may allow the scapula to seek conformity with the thorax by moving at the acromioclavicular joint in its other dimensions (ie, tilt and internal rotation). Scapular conformity on the thorax helps create a functional “gliding plane,” which may result in a more stable mechanism for performing upper extremity tasks.¹

During the final 2 phases of humerothoracic elevation (ie, 60 degrees–90 degrees and 90°–120°), acromioclavicular UR continues to serve as the predominant component motion. However, the proportional contribution decreases to an average of 42% during the 60- to 90-degree phase, and 37% during the 90- to 120-degree phase. This reduced acromioclavicular UR proportional contribution coincides with an increased sternoclavicular posterior rotation contribution, which now accounts for an average of 32% to 34% of scapulothoracic UR motion. The comparable contribution between acromioclavicular UR and sternoclavicular posterior rotation suggests that the scapula and clavicle now function as a mechanism, but do not fully become a single “claviscapular link,” as proposed by Dvir and Berme²³ and modeled by van der Helm.^{1  ,  2  ,  25} However, with the scapula and clavicle more coupled, the clavicle may be able to fulfill its role as a strut by supporting the upper limb on the thorax.³⁹ This function may be especially critical in higher phases of motion because of the higher dynamic loads between 60 and 120 degrees of humerothoracic elevation.⁴⁰

The extent to which the scapula and clavicle are coupled likely has implications on scapulothoracic muscle function. It has been generally described that the serratus anterior and middle and lower trapezius act at the sternoclavicular joints to produce net scapulothoracic UR and external rotation, respectively (Fig. 1).^{1  ,  41  ,  42} However, during the initial phase of humerothoracic elevation, the high acromioclavicular UR proportional contribution suggests that the scapula and clavicle are not yet coupled in most people. Therefore, contraction of the serratus anterior and middle and lower trapezius should primarily result in acromioclavicular joint motion. Although the moment arms for these muscles are large when acting at the acromioclavicular joint, they theoretically become much larger when acting at the sternoclavicular joint⁴³; this scenario can occur only when the clavicle and scapula move predominantly as 1 unit (ie, through sternoclavicular posterior rotation). An increased moment arm will result in increased torque producing capability, thus making the muscles more efficient at producing scapulothoracic UR and external rotation.

The increased proportional contribution from sternoclavicular posterior rotation during the 60- to 90-degree and 90- to 120-degree phases suggests that the serratus anterior acts, at least partially, at the sternoclavicular joint and may have increased leverage to produce net scapulothoracic UR and posterior tilt (Fig. 1). However, this increase in mechanical efficiency may also coincide with an increased requirement for the counterstabilizing action of the middle and lower trapezius to prevent the scapula from excessively internally rotating and laterally translating around the thorax in response to a more efficient serratus anterior.^{1  ,  41  ,  42  ,  44  ,  45} Therefore, it is possible that the degree to which the scapula and clavicle are coupled may influence the magnitude of scapular “dyskinesia” observed. For example, an individual in whom the scapula and clavicle are less coupled may require less trapezius and/or rhomboid muscle activation to maintain the position of the scapula on the thorax than someone in whom the scapula and clavicle are more coupled. The reason is that the serratus anterior may have a lower overall torque-generating capability while it still predominantly acts at the acromioclavicular joint in the individual in whom the scapula and clavicle are less coupled. The potential interaction between the magnitudes of coupling and muscle moment arms may help explain the often disparate findings reported in the literature regarding scapular kinematics in clinical populations^{6–11  ,  13  ,  46} and the presence of dyskinesia in both symptomatic and asymptomatic individuals.^{47  ,  48}

The relative proportional contribution between acromioclavicular UR and sternoclavicular posterior rotation may serve as a measure of “decoupling” in the shoulder complex. As such, this metric may have important implications for clinical decision making when examining patients with an acute acromioclavicular separation. Current guidelines for determining the appropriateness of surgical stabilization are based on the classification of Rockwood et al,⁴⁹ which relies upon static radiographs. However, this classification is limited by several factors that challenge its clinical utility, including radiographic projection error⁴⁹ and the inherent challenge of diagnosing a movement disorder using static images. Quantifying acromioclavicular UR and sternoclavicular posterior rotation proportional contributions may be a means to identify individuals who may benefit from surgical stabilization. Higher phases may be more sensitive to the presence of decoupling because scapular motion is at least partially transferred to the clavicle in most individuals without a history of acromioclavicular joint disruption (Fig. 5). Therefore, if an individual had an excessive acromioclavicular UR proportional contribution after 60 degrees of humerothoracic elevation, then the acromioclavicular and coracoclavicular ligaments may no longer effectively transmit motion, and surgical stabilization may deserve greater consideration. Future research is needed to rigorously define a cutoff for “excessive” acromioclavicular UR, as well as determine whether applying this cutoff improves surgical outcomes.

Throughout the range of motion, sternoclavicular elevation consistently accounted for < 11% of scapulothoracic UR. This low contribution suggests that the upper trapezius (through its insertion on the lateral clavicle⁴⁵) does not contribute substantially to scapulothoracic UR. Furthermore, the average acromioclavicular internal rotation angle was approximately 60 degrees, which according to the coupling function, suggests that only about 33% of sternoclavicular elevation will contribute to scapulothoracic UR. Consequently, increasing sternoclavicular elevation via shoulder “shrugging”—a movement impairment often observed clinically—may not be an efficient strategy to increase scapulothoracic UR.

The lack of significant interaction or group effects in proportional contributions suggests that shoulder complex coupling is not significantly affected in individuals with a diagnosis of impingement syndrome. However, it is important to emphasize that the degree to which the scapula and clavicle are coupled reflects the ability of the coracoclavicular ligaments to transfer motion and not necessarily the magnitude of motion. Because participants with a history of shoulder trauma or acromioclavicular joint separation were excluded from the study, any differences between groups would have indicated a physiologic disruption of acromioclavicular joint ligament function, which has not traditionally been associated with the impingement syndrome diagnosis. However, the heterogeneity of movement impairments in individuals with this clinical diagnosis may have impacted our ability to detect group differences.⁵⁰ Future studies investigating shoulder complex coupling may benefit from defining group membership on the basis of movement impairments, which would likely improve the statistical power and clinical interpretation of the results.

This study has limitations that should be considered. First, the shape and orientation of the clavicle relative to the imaging plane make shape matching challenging, especially at lower elevation angles. Initial validation work found clavicle axial rotation RMS errors of 3.7 degrees resulting in acromioclavicular UR errors of 3.4 degrees.²⁶ Therefore, shape-matching errors could explain why smaller sternoclavicular posterior rotation displacements were found compared to previous studies.^{5  ,  7} If sternoclavicular posterior rotation was in fact underestimated in this study, then the reported proportional contribution would be underestimated, while that of acromioclavicular UR would be overestimated.

Second, the derived coupling function assumes a linear relationship between scapulothoracic UR and its component motions. However, given trigonometric functions form the basis for kinematic calculations, it is likely a nonlinear function is required to better estimate proportional contributions. Despite the assumption of linearity, the derived function accounted for > 80% of component motion proportional contributions and its simplicity may facilitate clinical translation.

Third, this study only explored the kinematic mechanisms related to scapulothoracic UR and only during arm raising in the scapular plane. Other sternoclavicular, acromioclavicular, and scapulothoracic motions also serve important roles in shoulder complex function. Additional work is needed to investigate the kinematic mechanisms producing scapulothoracic anterior-posterior tilt and internal or external rotation as scapular dyskinesias are often seen in combination.⁵¹ Future studies are also needed to translate this kinematic knowledge into kinetics in which muscle function—and, potentially, dysfunction—can be studied, which can lead to evidence-based exercise recommendations.

Finally, shoulder complex kinematics were investigated using sophisticated research technology that is not yet clinically translatable. However, this study adds to the foundational knowledge regarding the coupling kinematics of scapulothoracic motion, which is required in order to interpret impaired kinematics within clinical populations.

Despite the complex anatomical and kinematic relationships of the shoulder complex, scapulothoracic UR can be effectively predicted from the component motions as a function of acromioclavicular internal rotation. Acromioclavicular UR and sternoclavicular posterior rotation are the predominant component motions of scapulothoracic UR with a combined contribution of at least 70%. More research is needed to investigate how these coupling relationships influence muscle function and scapular dyskinesia.

Author Contributions and Acknowledgments

Concept/idea/research design: R.L. Lawrence, J.P. Braman, P.M. Ludewig

Writing: R.L. Lawrence

Data collection: R.L. Lawrence

Data analysis: R.L. Lawrence, D.F. Keefe, P.M. Ludewig

Project management: R.L. Lawrence, P.M. Ludewig

Fund procurement: R.L. Lawrence, P.M. Ludewig

Providing participants: R.L. Lawrence

Providing facilities/equipment: J.P. Braman, P.M. Ludewig

Providing institutional liaisons: P.M. Ludewig

Consultation (including review of manuscript before submitting): R.L. Lawrence, J.P. Braman, D.F. Keefe, P.M. Ludewig

The authors thank the study participants and the radiology technologists who assisted with magnetic resonance and fluoroscopic image acquisition: Wendy Elvendahl, Cassi Koldenhoven, Erik Solheid, and Scott Haglund.

Ethics Approval

The study was approved by the Institutional Review Board and All-University Radiation Protection Advisory Committee at the University of Minnesota. All participants provided written informed consent.

Funding

Research reported in this article was supported by the Eunice Kennedy Shriver National Institute of Child Health and Human Development (F31-HD087069 and L30-HD089226), the National Institute of Arthritis and Musculoskeletal and Skin Diseases (T32-AR050938), the National Center for Advancing Translational Sciences (UL1-TR002494), and the National Institute of Biomedical Imaging and Bioengineering (P41-EB015894). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. This work was also supported in part by Promotion of Doctoral Studies (PODS) Scholarships from the Foundation for Physical Therapy, the Minnesota Partnership for Biotechnology and Medical Genomics, a research infrastructure grant from the University of Minnesota Office of the Vice President for Research, and the University of Minnesota Department of Orthopaedic Surgery and Clinical and Translational Science Institute.

Disclosures and Presentations

The authors completed the ICMJE Form for Disclosure of Potential Conflicts of Interest and reported no conflicts of interest.

This work was completed while R.L. Lawrence was enrolled as a doctoral candidate in the Division of Rehabilitation Science, Department of Rehabilitation Medicine, University of Minnesota. The work should be fully attributed to the University of Minnesota.