Evidence-Based Strength Training: Scapulothoracic Musculature, Part 2

Scapulothoracic Muscles and Pain

As I mentioned in Part 1, weakness or poor neuromuscular control of the periscapular muscles has been implicated in subacromial impingement1,2, lateral epicondylalgia3-5, cervicogenic headache6, and neck pain7,8.

Specifically, insidious onset of neck pain and whiplash associated disorder (WAD) have been linked with a significant delay in and shorter duration of serratus anterior activity bilaterally during arm elevation9. A similar study found decreased serrates anterior activation in individuals with acromioclavicular osteoarthritis and rotator cuff disease.10

Although a cause-and-effect relationship cannot be confirmed, this preliminary evidence still lends support for targeting the periscapular muscles in individuals with neck or shoulder pain.

EMG Activity and Exercise Goals

According to Reiman et al.11 and Escamilla et al.12, moderate EMG activation (21-40% MVIC) is best used to facilitate endurance and neuromuscular re-education; high activation (41-60+% MVIC) – to promote strength gains.

From Biomechanics to Exercises

Serratus Anterior

Primary Function: scapular upward rotation, external rotation, posterior tilt at the acromioclavicular joint, protraction of the clavicle at the sternoclavicular joint.

Origin: External surfaces of lateral aspect of 1st-8th ribs

Insertion: Anterior surface of medial border of scapula

The SA is often activated with scapular protraction. The exercises yielding the highest MVIC for the serratus anterior include:

  • dynamic hug13
  • push-up plus13
  • scaption with external rotation14
  • diagonal PNF (shoulder flexion, horizontal flexion, external rotation)15
  • shoulder abduction in scapular plane above 120 degrees15

The upper trapezius (UT) often compensates for a weak or inhibited serratus anterior, so it’s beneficial to selectively activate the SA in lieu of the UT. According to Cools and colleagues, the best SA:UT ratio is achieved in:

  • high row
  • forward shoulder flexion
  • scaption with external rotation14
Serratus Anterior
Prone Push Up with Plus
Serratus Anterior
Dynamic Hug with Resistance
[Table] MVIC values for SA exercises

Levator Scapulae

Primary Functions: scapular elevation, glenoid cavity inferior tilt through upward scapular rotation

Origin: Posterior tubercles of transverse processes of C1-C4 vertebrae

Insertion: Medial border of scapula superior to root of scapular spine

These muscles have received little attention in the literature compared the SA or trapezius. In a study, Moseley and colleagues discovered that the levator scapulae achieves the highest activity in:

  • rowing
  • horizontal abduction
  • shrug
  • horizontal abduction with ER
  • prone shoulder extension16
[Table] MVIC values for Levator Scapulae exercises


Primary Functions: Retraction of the scapula; upward rotation to depress glenoid cavity; scapular attachment to thoracic wall

Origin: nuchal ligament; spinous processes of C7, T1 and T2-T5 vertebrae

Insertion: smooth triangular area at medial end of scapular spine; medial border of scapula from level of spine to inferior angle

The rhomboids achieve the highest MVIC during:

  • ER at 90° of abduction17
  • ER at 0° of abduction17
  • horizontal abduction16
  • shoulder extension17
  • scaption16
Levator Scapulae
Prone Shoulder Row
Shoulder External Rotation at 90° Abduction with Dumbbell
[Table] MVIC values for rhomboid muscle exercises

Choosing the Best Exercise

These studies give us a glimpse into properly selecting exercises, yet very few exercises have been or will ever be studied.

When choosing an exercise for your patient, be sure to consider:

  • the biomechanics of the movement,
  • current evidence for or against the exercise,
  • your patient’s presentation, and goals for treatment.
  • aggravating movement(s) or comparable signs

Evidence-Based Strength Training: Scapulothoracic Musculature, Part 1

In the next installment of the Evidence-Based Strength Training Series for MedBridge Education, we are going to take a look at the often-neglected scapulothoracic musculature. Typically when considering the management of painful upper quarter conditions, local exercise and manual therapy interventions are employed judiciously. However, when utilizing a proper movement assessment or regional interdependence philosophy, impairments in the scapulothoracic musculature are often found to contribute to pain in distal or proximal joints. Weakness or diminished neuromuscular control of the peri-scapular muscles has been implicated in Subacromial Impingement4,18, Lateral Epicondylalgia2,7,12, Cervicogenic Headache10, and Neck Pain3,16. Additionally, in a prospective cohort study conducted by Clarsen et al.6, the presence of scapular dyskinesis led to an 8.4x greater risk for developing a shoulder injury during the course of an elite male handball season. Furthering the support for interventions focusing on the peri-scapular musculature, Lawrence and colleagues11 found that in the presence of shoulder pain due to subacromial impingement, patients demonstrated significantly reduced scapulothoracic upward rotation at lower angles of humerothoracic elevation and significantly reduced sternoclavicular posterior rotation throughout humerothoracic elevation.

With the knowledge of the scapulothoracic musculature’s impact on potentially injurious altered biomechanics and the deficits seen in many common musculoskeletal disorders, healthcare providers need to ask themselves, “What are the best exercises to recruit these muscles?”

Before delving into specific exercises, it is necessary to understand the basic biomechanics of the scapulothoracic and glenohumeral joints. During humeral elevation, the scapula upwardly rotates 1° for every 2° of humeral elevation until 120° humeral elevation is achieved. After this point, scapular rotation contributes 1° for every 1° humeral elevation until maximal arm elevation is met. Also, the scapula typically tilts posteriorly between 20° and 40° in the sagittal plane and externally rotates between 15° and 35° in the transverse plane17. This complex movement pattern relies on coordinated and balanced contributions from the trapezius, serratus anterior, levator scapulae, rhomboid, and pectoralis minor musculature.

Trapezius Musculature

The broad posterior musculature known as the trapezius originates at the medial third of superior nuchal line, external occipital protuberance, nuchal ligament, and the spinous processes of C7-T12 vertebrae and its distal insertion is at the lateral third of clavicle, acromion process, and spine of scapula. This muscle is divided into three distinct portions with the Upper Trapezius (UT) providing scapular elevation, Lower Trapezius (LT) proving depression, and the Middle Trapezius (MT) causing scapular retraction. Additionally, the UT and LT act together to rotate the glenoid cavity superiorly, which is a very important and often dysfunctional action for individuals suffering from shoulder impingement or pain11.

The primary action of the upper trapezius musculature involves elevation of the scapula and, predictably, the exercises that provide the highest Maximal Isometric Voluntary Contraction (MVIC) are those that involve this motion. Additionally, during scapular abduction, upper trapezius activity progressively increases from 0° to 60° and from 120° to 180° of abduction1. With this knowledge in mind, researchers have found that the highest electromyographical (EMG) activity occurs during the uni-lateral shoulder shrug9, rowing14, scaption8, and shoulder abduction in the scapular plane above 120°9. Due to the infrequency of UT weakness (unless secondary to neurological involvement) strengthening of this portion of the trapezius is often not the focus during the rehabilitation of painful upper quarter conditions. Instead, clinicians have learned to focus on improving middle and lower trapezius strength and normalizing the ratio of UT to the lower two portions of the trapezius activation.

[Table] Click to see MVIC values for UT exercises

Similar to the UT, the middle trapezius, with its primary function being scapular retraction, is often activated during exercises involving this action. The highest MVIC percentages for the MT have been recorded during horizontal abduction14, prone full-can9, horizontal abduction with external rotation14, and scaption8. Additionally, as the UT often compensates for a weak MT or lower trapezius, it is likely beneficial to utilize exercises with a good upper trapezius to middle trapezius ratio (UT:MT) when attempting to strengthen this musculature. Exercises shown to provide this ratio are side-lying forward flexion, side-lying external rotation, and prone shoulder extension5.

[Table] Click to see MVIC values for MT exercises


Due to its impact on scapular upward rotation, external rotation, and posterior tilt, the lower trapezius is of more importance than the aforementioned UT and MT during rehabilitation17. There have been a multitude of studies investigating lower trapezius weakness and its association with painful conditions and most do find this connection. Due to this fact, there have also been many studies looking into maximal EMG activity of the lower trapezius during upper extremity strengthening. The results show that significantly high MVIC values have been recorded during arm raise overhead in line with the LT muscle fibers9, ER at 90° of abduction9,15, horizontal abduction with ER5, and prone shoulder abduction5. While choosing exercises with a relatively high MVIC is important, it may be more important to choose those exercises that provide an optimal upper trapezius to lower trapezius ratio (UT:LT). As these are the two primary muscle groups involved in upward rotation of the scapula, having adequate and relatively equal contributions from each is important in maintaining normal biomechanics. In a study conducted by Cools and colleagues, it was determined that side-lying forward flexion, side-lying external rotation, and horizontal abduction with external rotation had the best UT:LT ratios5. In addition to this study, McCabe et al. conducted a similar study and found that the seated press-up, uni-lateral scapular retraction, and bilateral shoulder external rotation provided UT:LT ratios that showed a preferential activation of the lower trapezius over the upper trapezius13. While these studies do give a glimpse into proper exercise selection, not every exercise has been studied to date and never will. When choosing an exercise for your patient/client, it is important to take into consideration the biomechanics of the movement, current evidence supporting/refuting, and your patient’s presentation and goals for treatment.

[Table] Click to see MVIC values for LT exercises


The trapezius musculature is a very important piece of the puzzle, but contributions from the Serratus Anterior, Rhomboids, and Levator Scapulae also play a large role and will be discussed in Part 2.


Differential Diagnosis: Superior Labral Anterior-Posterior (SLAP) Lesions

The following is another article written for the online, video-based physical therapy continuing education company MedBridge

Amongst overhead throwing athletes, there are several injuries that typically come to mind, and at the top of that list is the Superior Labral Anterior to Posterior (SLAP) Lesion. While this is not an overly common injury for the general population, with an overall incidence of 26% (Kim et al, 2003), diagnosis and subsequent surgical intervention does seem to be on the rise. Onyekwelu et al reported that the incidence of SLAP repairs in New York State rose by a factor of 5.5 between 2002 and 2010. Similarly, Weber et al analyzed 4,975 cases of SLAP repairs collected from the American Board of Orthopaedic Surgery (ABOS) and found that ABOS candidates were performing SLAP repairs at a rate 3 times higher than expected. With the growing incidence of this pathology, clinicians must understand which mechanism of injury and patient presentation is indicative of a SLAP lesion.

Mechanism of Injury

The proposed mechanism of injury is far from definitive with several theories available. Originally, Andrews et al theorized that eccentric load to the biceps during throw deceleration resulted in a tension-distraction traumatic or repetitive injury to the biceps-labral complex. Whereas, Snyder et al determined that a fall on the outstretched arm was the most common mechanism of injury in their patients and believed a combination of a compression force to the superior joint surface and proximal subluxation force on the humeral head was the mechanism of injury. Additionally, Walch et al and then Jobe et al proposed a separate theory which involved the concept of internal impingement as a cause of pathology to the rotator cuff and labrum in the throwing athlete.

This internal impingement is caused when the shoulder is excessively abducted and externally rotated, creating an impingement of the postero-superior glenoid and undersurface of the rotator cuff. The most recent theory was proposed by Burkhart et al in 1998 and involves a ‘peel back’ mechanism of the biceps-labral complex. The authors suggest that when the shoulder is placed in a position of abduction and maximal external rotation, a twist results at the base of the biceps where a torsional force is then transmitted. This theory seems to hold up in cadaveric studies as Burkhart’s peel-back mechanism created a significantly weaker biceps anchor when compared to Andrew’s eccentric load during deceleration (Shepard et al, 2004). Amongst these four proposed mechanism of injury, the peel-back mechanism seems to be the most plausible, however all theories should be taken into consideration when determining the presence of a potential SLAP lesion.

Patient Symptoms

Patients presenting with a SLAP lesion typically have several very similar complaints that led to them seeking medical attention. The patient usually presents with a nonspecific history of shoulder pain or “giving way” of the shoulder with overhead activities, although in some cases there can be a history of a traumatic injury. The patient will complain of anterior shoulder pain with ‘clicking’ and/or ‘popping’ sensations during the late cocking phase or early acceleration phases of throwing (Barber et al, 2008). Coinciding with these traits, patients may complain of a gradual loss of throwing velocity and/or control (Bedi et al, 2008). In agreement, Snyder et al found that 50% of overhead athletes presenting with a SLAP lesion report a loss of velocity and accuracy along with general uneasiness of the shoulder. Furthermore, older patients may complain of generalized shoulder weakness secondary to rotator cuff involvement, whereas younger patients are more likely to present with symptoms of instability. In addition to these complaints, many overhead athletes may develop a loss of glenohumeral internal rotation with a subsequent increase in external rotation as an adaptation to the demands of their sport. This is known as Glenohumeral Internal Rotation Deficit (GIRD) and allows for increased velocity, but also predisposes them to labral pathology based on the theorized mechanism of injury proposed by Walch et al.

Testing for SLAP Lesions

Unfortunately, with regards to special testing for SLAP lesions, there are very few tests that show any statistical benefit. A recent systematic review with meta-analysis conducted by Hegedus and colleagues found that none of the 8 special tests available for meta-analysis demonstrated sensitivity, specificity, or likelihood ratio values capable of ruling in or out the presence of a SLAP lesion. The likelihood ratios for the included special tests indicated little more than a 50/50 chance in diagnosing this pathology with positive likelihood ratios ranging from 0.90 to 2.86 and negative likelihood ratios from 0.87 to 1.03. Of particular interest, the popular O’Brien’s Active Compression test had very poor findings (Sn= 0.67, Sp= 0.37, +LR= 1.06, -LR= 0.89) when the original study was removed from the analysis due to the drastically different findings reported in O’Brien et al’s initial investigation compared to the six more recent studies. In light of this information, in his course titled, “Evidence-Based Examination of the Shoulder”, Eric Hegedus suggests the use of four completely different and more statistically relevant special tests:

While these tests are more recent and have not had a large amount of research to evaluate their diagnostic ability, they offer far and away better statistical probability of an accurate diagnosis compared to other more common tests used in clinical practice.

Thorough Examination and Accurate Diagnosis

Finally, it should also be noted that SLAP lesions are rarely a condition in isolation. Andrews et al reported 45% of patients (73% of baseball pitchers) with SLAP lesions had concomitant partial thickness tears of the supraspinatus portion of the rotator cuff. While Mileski and Snyder reported that 29% of their patients with SLAP lesions exhibited partial thickness rotator cuff tears, 11% complete rotator cuff tears, and 22% Bankart lesions of the anterior glenoid. This indicates that just because another more common pathology is diagnosed, it does not mean that a SLAP lesion can be ruled out. A thorough examination involving the testing of all likely differential diagnoses must be conducted for those patients presenting with an overhead throwing injury.

While diagnosis of the Superior Labral Anterior to Posterior lesion is far from an easy or straightforward process, there are aspects that can be used to improve the likelihood of an accurate diagnosis. Patient characteristics are of utmost importance with overhead throwing athletes complaining of decrease in velocity, anterior shoulder pain, and the presence of GIRD. Additionally, a change needs to be made with regards to special testing of the shoulder away from O’Brien’s Active Compression Test to the more statistically favorable passive distraction, passive compression, and both versions of the dynamic labral shear test. As research evolves, we as clinicians must do the same in order to improve our ability to accurately diagnose and treat the complex pathologies present in the overhead throwing athlete.

When Can I Play Again? Return to Sports Testing for the Upper Extremity.

The following is another article written for the online, video-based physical therapy continuing education company MedBridge

A lot has been written and researched with regards to return to sport criteria and testing for injuries of the lower extremity, and more specifically following anterior cruciate ligament reconstruction (ACL-R), however little attention has been given to injuries of the upper extremity. As with ACL-R, return to sport following surgical intervention in the upper extremity is less than stellar. Harris et al conducted a systematic review that found amongst elite pitchers undergoing shoulder surgery (rotator cuff, biceps/labrum, instability, internal impingement, ect.), only 68% returned to play 12 months following surgery. Additionally, they found that 22% of major league baseball pitchers included in their review never returned to sport. In agreement with these findings, Cohen et al evaluated the return to sport of professional baseball players following shoulder and/or elbow surgery and found only 48%  of participants returned to the same or higher level of professional baseball following surgery. Why are these numbers so low and what can we do as rehabilitation specialists to improve the rate of return to sport following surgery?

Sometimes, it simply takes correctly identifying those who are at risk of re-injury or those simply not ready to rerun to their chosen sport. When devising an appropriate return to sport test, Phil Plisky, PT, DSc, OCS, ATC, CSCS says in his course, “Return to Sport and Discharge Testing“, that each test should be reliable, predictive of injury, have discriminate validity, and the test must be modifiable with training/rehabilitation. With regards to the upper extremity, there is a significant gap in knowledge/research in comparison to the lower extremity. That being said, the Y-Balance Test has recently been adapted to help fill this gap. Gorman et al investigated to reliability of the Upper Quarter Y Balance Test (UQ-YBT) and found that the test-retest reliability (0.80-0.99) and inter-rater reliability (1.00) ranged from good to excellent. Along with this information, normative data was determined amongst active adults with males generally performing the test superiorly to females and a minimally detectable difference of 8.1 cm in the medial direction, 6.4 cm in the superolateral direction, and 6.1 cm in the inferolateral direction. In addition to these findings, Westrick et al found that there was no significant difference between the dominant and non-dominant limb when young females or males perform the UQ-YBT. This shows that, generally speaking, any significantly asymmetrical findings should be investigated further prior to returning the athlete to his/her sport. While, currently, there are no studies investigating this test’s capacity to predict injury or its ability to be modified with training, the excellent reliability and discriminate validity make this a solid return to sport test.

Similarly, the Closed Kinetic Chain Upper Extremity Stability Test (CKCUEST) offers an additional way to assess upper extremity dynamic stability, albeit in a singular plane. Once again, this test demonstrates excellent reliability with a Test-Retest Reliability of 0.92 (Goldbeck et al), an intersession reliability ranging between 0.87 to 0.96 (Tucci et al), and an intrasession reliability ranging between 0.86 and o.97. Furthermore, Tucci et al also found the CKCUEST to have discriminate validity as those performing the test with diagnosed subacromial impingement performed significantly inferiorly in comparison to asymptomatic participants. Along with this excellent reliability and obvious display of closed kinetic chain dynamic stability, the CKCUEST also has recently been shown to have the capacity to predict injury. Pontillo et al performed a prospective cohort study attempting to identify potentially factors that would be predictive of upper extremity injury in collegiate football players. The only significant factor in predicting future injury in this population of athletes was a CKCUEST in which the athlete completed < 21 touches (Sn= 79%, Sp= 83%, + LR= 4.74, – LR= 0.25, Odds Ratio= 18.75). This is a significant finding and shows the benefit for utilizing this test not only for return to sport, but also in pre-season testing to identify individuals who are at risk for injury.

For a more demanding task, similar to the single-leg hop testing utilized for patients following ACL reconstruction, the One-Arm Hop Test was created to test the athlete’s plyometric, power, and dynamic closed kinetic chain stability. Unfortunately, to this date, there has only been one study investigating this specific return to sport test. Falsone et al found the test to have good Test-Retest Reliability (0.78-0.81) and also found only a 4.4% difference between non-dominant and dominant limbs when performing the test. This once again shows the ability to assess post-operative function based upon the symmetry between limbs. While this may not be a perfect solution, it allows the ability to utilize the test with evidence-based backing until further research is conducted investigating its ability to predict injury and/or be modified with training.

Returning an athlete to sport is a multi-factorial decision that must incorporate that athlete’s psychological readiness to return to play, strength, range of motion, pain level, and ultimately the ability to perform the movement patterns consistent with their sport and/or position. The aforementioned return to sport tests provides a hierarchical (i.e. increasingly demanding) system for testing the individual’s capacity to withstand the rigors of their chosen activity. This allows clinicians something outside of subjective reports, range of motion, and strength measures to assess your patient’s ability to perform dynamic upper extremity tasks prior to returning to sport and in doing so, we may be able to identify some of the deficits our athletes are hiding that are preventing them from ultimately returning to their sport.

NMES, The Missing Link.

The following is another article written for the online, video-based physical therapy continuing education company MedBridge Education

Following any type of surgery, significant weakness of the primary and secondary musculature is common. For example, quadriceps weakness has been documented during the immediate post-operative phase following surgery (Snyder-Mackler et al), as well as years following rehabilitation (Rosenberg et al). Additionally, patients who undergo Total Knee Arthroplasty (TKA) exhibit similar findings. According to Mizner et al and Stevens et al, quadriceps strength drops 50-60% of pre-operative levels one month following TKA, despite the initiation of rehabilitation within 48 hours of surgery. Following this trend, Rokito et al found external rotation deficits following rotator cuff repair of 79% and 90% at six months and one year, respectively. Considering the severity and chronicity of these strength deficits, more effective interventions are warranted to restore strength and improve long-term outcomes. One particular modality that has been shown to improve these deficits is neuromuscular electrical stimulation (NMES).

Kim et al recently published a systematic review investigating the utility of NMES following ACL reconstruction to improve quadriceps function and strength. In this review, which involved 8 randomized controlled trials (RCTs), it was demonstrated that NMES in conjunction with exercise, compared to exercise alone or in combination with electromyographic bio-feedback, results in greater quadriceps strength recovery. As discussed in a previous article, return to sport is the ultimate goal for most patients and quadriceps femoris strength is of the utmost importance. Schmitt et al conducted a cross-sectional study to determine the impact of quadriceps weakness on return to sport functional testing. Those patients who presented with a quadriceps index (quadriceps strength involved/uninvolved) of < 85% performed inferiorly when compared to those with a quadriceps index of > 90%. Additionally, quadriceps weakness predicted performance on single-leg hop testing regardless of graft type, presence of meniscus injury, knee pain, and knee symptoms. Similarly, Fitzgerald et al not only measured increased quadriceps strength, but also length of time until progression to agility/plyometric training.  This randomized controlled trial found that those in the NMES group met the criteria for progression in a greater proportion than those in the control group. At 16 weeks, 85.7% (18/21) of patients receiving NMES achieved progression to agility training, whereas only 68% (15/22) of those in the control group achieved similar results.

In addition to ACL rehabilitation, those undergoing total knee arthroplasty enjoyed similar benefits. Stevens-Lapsley et al conducted a prospective, longitudinal, randomized controlled trial investigating the effects of NMES on patients following TKA. Patients were randomized into a group receiving standardized rehabilitation or to a group receiving the same rehabilitation in addition to NMES, which was initiated 48 hours following surgery. In both the short-term (3.5 weeks) and long-term (52 weeks), patients in the NMES group demonstrated superior quadriceps strength, hamstring strength, and functional performance (Timed “Up & Go” Test, the Stair-Climbing Test, and the Six-Minute Walk Test). Additionally, Walls et al investigated the pre-operative utility of this modality. Those individuals who received NMES achieved significant increases in quadriceps strength from weeks 6-12, whereas the control group did not achieve the same feat. Lastly, in a case report published by Petterson et al, a cyclist presenting 12 months following bilateral TKA displayed significant impairments with regards to quadriceps strength and volitional muscular contraction. Following six weeks of NMES and volitional therapeutic exercise, this patient achieved a 25% improvement in left quadriceps femoris maximal volitional force output and his central activation ratio (CAR) also improved from 0.83 to 0.97 as quantified by the burst superimposition technique. Furthermore, strength gains continued after the end of treatment as his quadriceps strength index was 94% of his right leg at 12 months following treatment.

While the majority of research pertaining to the efficacy of NMES has been done in the lower extremity, this is not the only region where its benefits can be found. As previously stated, muscular deficits frequently accompany patients following rotator cuff repair. To this end, Reinold et al investigated the ability of NMES to affect maximum voluntary contraction of the infraspinatus muscle. In comparison to trials without NMES, peak force production was significantly greater with an average force of 3.75 kg in comparison to just 3.08 kg. This increase was present regardless of patient age, size of the tear, intensity of the current, or the number of days following surgery. While this preliminary study does not give credence to the effect during a full course of rehabilitation, it does speak to the ability of NMES to increase the ability of this musculature to contract safely and more efficiently following surgery. Further research will define the effectiveness of this intervention following rotator cuff pathology, however this study lends hopeful possibilities.

Neuromuscular electrical stimulation should play an integral role in your practice regardless of setting. Patients presenting with strength deficits and impairments will benefit from NMES when combined with volitional exercise. Meryl Gersh, PT, PhD goes into great detail with regards to electrode placement, optimal dosage, and indication criteria in her course “Applying Electrical Stimulation in Your Physical Therapy Practice”. Increasing your patients’ volitional muscular contraction is of the utmost importance when it comes to fostering improved long term outcomes and NMES in conjunction with their current program should yield enhanced results.

Evidence-Based Strength Training: Rotator Cuff

This will be the first in a series of monthly posts that I will be contributing to MedBridge Education, who is an online continuing education resource for physical and occupation therapists…

According to Sipes et al, 30% of athletes suffer a shoulder injury during their career. Of those injuries, subacromial impingement syndrome and rotator cuff tendonitis were the most common shoulder injuries for each individual sport and accounted for 27% and 24% of total shoulder injuries, respectively. More specifically, over-head throwing athletes are especially susceptible to shoulder pathology as 28% of all injuries in professional baseball have been shown to occur at the shoulder (Cante et al).

Additionally, van der Windt et al conducted a prospective evaluation of over 300 patients. This analysis found that approximately 48% of shoulder injuries in a general population were diagnosed as subacromial impingement. Stability at the glenohumeral joint is dependent upon the passive stabilization of the ligamentous tissues and the dynamic stability provided by the rotator cuff musculature, as well as indirectly by the stability of the scapula and the muscles that support it. The rotator cuff’s primary role is to center the humeral head in the glenoid fossa, which requires adequate muscular strength and endurance. Because the rotator cuff muscles all originate from the scapula, appropriate functional control of the scapula is also an important rehab target.

As weakness of the rotator cuff and scapular stabilizing musculature can predispose individuals to subsequent pathological conditions, understanding what exercises are most effective in training these muscles is of utmost importance to the practicing clinician and student alike.  Evidence supports the use of appropriate exercises and manual therapy to improve the strength and function of these muscles. For example, in a clinical trial (Bang et al) comparing an exercise program to the same program with the addition of manual therapy, strength of patients in the manual therapy program improved significantly whereas the exercise-only group did not. In addition, based on the unique ROM, capsular laxity, strength, proprioception, and osseous anomalies specific to the over-head athlete, they too require the creation of a rehabilitation program specific to the demands of their sport/position (see: “Rehabilitation of the Overhead Throwing Athlete”). This review will touch on strength training interventions targeting the rotator cuff, but please remember this is only the tip of the iceberg in terms of rehabilitative principles.


The supraspinatus originates from the supraspinous fossa and inserts laterally to the superior facet of the greater tubercle of the humerus. Based on this muscle’s line of pull, its primary responsibility involves concentric abduction of the humerus. In addition to this primary action, the supraspinatus also drives the superior roll of the humeral head and compresses the humeral head firmly against the glenoid fossa during shoulder abduction. Secondarily, this muscle contributes a small external rotation torque. Supraspinatus mm Based on these biomechanical and anatomical considerations, Reinold et al investigated the electromographic (EMG) activity of the supraspinatus through  three common therapeutic exercises, which also serve as diagnostic tests for shoulder impingement when performed isometrically (Full Can, Empty Can , and Prone Full Can). All three of these exercises had nearly identical EMG data ranging between 62% (Full Can) and 67% Maximal Voluntary Isometric Contraction (MVIC) (Prone Full Can). It should also be noted that there was significantly greater middle and posterior deltoid EMG activity during the empty can exercise, which can contribute to superior humeral head migration and a predisposition to subacromial impingement. Biomechanically speaking, when the humerus is elevated in a position of internal rotation, it does not allow the greater tuberosity to clear from under the acromion as it does in a neutral or externally rotated position. Based on these findings, why are we still using the more provocative Empty Can when the better tolerated full can exercise is just as effective? The answer for this is still unclear.

Earlier in 2004, Reinold et al conducted a similar study looking at a broader range of exercises and muscular contributions of the supraspinatus, infraspinatus, teres minor, and posterior & middle deltoid. It was determined that the top three exercises based on %MVIC were Prone Horizontal Abduction at 100° with full external rotation (82% MVIC), Prone External Rotation at 90° of abduction (68% MVIC), and Standing External Rotation at 90° of abduction (57% MVIC). As previously stated, selecting exercises that activate the supraspinatus while minimizing the activity of the deltoid musculature is of importance in the rehabilitation of most shoulder pathologies. This information could make utilizing the Prone Horizontal Abduction at 100° with Full External Rotation and Prone External Rotation at 90° of abduction detrimental to proper rehabilitation as substantial deltoid activity was recorded at 82% and 79% MVIC, respectively.

In lieu of this additional consideration, the most appropriate exercises for isolated strengthening of the supraspinatus are the Full Can, Prone Full Can, and Side-lying ER (51% MVIC). Once progressing to a more functional program targeting scapular and glenohumeral stability, it must be noted that this does not necessarily warrant the discharge of more targeted rotator cuff exercises in favor of seemingly more difficult weight-bearing exercises. Uhl et al evaluated several common weight-bearing exercises and of the exercises evaluated, the most intensely the supraspinatus was engaged was only 29% MVIC during a single-arm push-up.


The infraspinatus originates distally to the supraspinatus at the infraspinous fossa of the scapula and attaches laterally at the middle facet of greater tubercle of the humerus. Based on this muscle’s origin and insertion, its primary action involves external rotation of the humerus. Infraspinatus mm. The secondary actions of the infraspinatus include horizontal abduction, glenohumeral compression at the glenoid fossa, and resistance to superior and anterior humeral head translation. This is why many diagnostic special tests for impingement or rotator cuff integrity will also assess the strength of the infraspinatus (e.g. the Resisted Isometric External Rotation Test). The previously mentioned study conducted by Reinold and colleagues also evaluated the %MVIC of the infraspinatus musculature during the same exercises. The three most demanding exercises for the infraspinatus included Sidelying External Rotation at 0° of abduction (62% MVIC), Standing External Rotation in the scapular plane (53% MVIC), and Prone External Rotation at 90° of abduction (50% MVIC). Based on the line of pull and perceived action of the infraspinatus, this data seems to make reasonable sense. Additionally, when having your patient perform side-lying external rotation, consider having them place a rolled towel between their arm and their torso. By making this simple adjustment, it is postulated that the muscles controlling adduction and those performing external rotation are more appropriately balanced. The data from these two exercises agree with this theory as %MVIC of the infraspinatus increased from 20% to 25% MVIC with the addition of a rolled towel. When progressing to more functional weight bearing exercises, Uhl et al determined that infraspinatus activity was substantially more active than the other rotator cuff musculature measured. They found the Push-up with feet elevated (52% MVIC) and One-armed Push-up (82% MVIC) to be especially challenging. Obviously, these are more challenging activities and are not appropriate for every patient, but may be beneficial adjunctive exercises for the more advanced clientele.

Teres Minor:

The smaller, but still important, teres minor attaches inferior to the infraspinatus at the lateral border of the scapula and inserts onto the inferior facet of the greater tubercle of the humerus. The teres minor’s primary role is external rotation and stabilization of the humeral head within the glenoid fossa. Secondary actions include adduction and horizontal abduction of the humerus. Teres Minor mm. While, in general, teres minor performs similar actions to the infraspinatus. It provides drastically less activity during flexion, abduction, and scapular abduction than the infraspinatus. It may be isolated with the Dropping Sign Test, which has been shown to have poor sensitivity, but good specificity (Hertel et al).  With regards to exercise prescription, the three most demanding exercises (as determined by Reinold and colleagues) are Sidelying External Rotation at 0° of abduction (67% MVIC), Standing External Rotation in the scapular plane (55% MVIC), and Prone External Rotation at 90° of abduction (48% MVIC). Once again, given this muscle’s line of pull, the data with regards to these exercises make both biomechanical and anatomical sense. Horizontal abduction was also evaluated as it was shown to have substantial activation (74% MVIC) in a previous study conducted by Townsend et al. However, Reinold’s more recent study showed a relatively small contraction of both the infraspinatus (39% MVIC) and teres minor (44% MVIC). As both the infraspinatus and teres minor are the primary external rotators at the glenohumeral joint, it is important to understand which exercisesimultaneously activates both muscles. Reinold has determined the exercises that elicite the highest combined EMG activation are shoulder ER in side-lying, standing ER in the scapular plane at 45° of abduction, and prone ER in 90° of abduction.


The subscapularis is the only rotator cuff muscle that is located on the anterior surface of the scapula. With an origin at the subscapular fossa and insertion laterally onto the lesser tubercle of the humerus, its primary action is internal rotation of the humerus. Secondary actions include humeral adduction, production of an abduction torque during arm elevation, glenohumeral compression, and anterior stabilization of the glenohumeral joint. Subscapularis mm. As with the supraspinatus, the subscapularis can produce its maximal force when the humerus is positioned at 0° of abduction. As the abduction angle increases, the moment arms of the inferior and middle heads stay relatively constant.However, the moment arm of the superior head progressively decreases until approximately 60° abduction, which translates into diminished torque production. There have been many studies with conflicting results in terms of optimal abduction angle for subscapularis force production. In place of definitive EMG conclusions, potential for compensation should be taken into consideration. With the arm positioned at 0° of abduction, Decker et al found increased activation of the pectoralis major, latissimus dorsi, and teres major, which indicates a greater potential for substitution and masking of subscapularis weakness. In contrast, it was determined that pectoralis major activity decreased substantially when internal rotation was performed at 90° of abduction.

While this study did not clear up any of the murkiness in regards to subscapularis activation, it did offer assistance in avoiding or detecting compensatory muscular substitution. Decker et al also determined three more complex movements that created substantial subscapularis activity. The Push-up Plus (135.5% MVIC), Diagonal (99.7% MVIC), and Dynamic Hug (94.1% MVIC) exercises all created subscapularis activity that exceeded performance during isolated internal rotation exercises. Although these values are much greater than those reported for isolated internal rotation exercises, the potential for muscular substitution is likely with these more dynamic and functional exercises. Based on these considerations, isolated internal rotation should not be dismissed, but rather used in conjunction with these more demanding activities.

Additional Considerations:

Movement at the glenohumeral joint is complex and dependent on many different muscular actions as well as the contributions of several other joints. While strengthening of the rotator cuff and scapular stabilizers is often directed for patients with shoulder pathology, other regions and joints must also be evaluated. Contributions of the cervical spine, thoracic spine , scapulothoracic joint, acromioclavicular joint , sternoclavicular joint, and passive ligamenous structures can affect the underlying pathology and subsequent rehabilitation process. This is why manual therapy is often directed at these areas. Not only can it increase  mobility, but it can also improve the quality and strength of the exercise (Bang et al).

Additionally, scapulothoracic and pectoral musculature is also necessary for optimal shoulder biomechanics and must be addressed when appropriate. A recent continuing education course taught by Lenny Macrina, MSPT, SCS, CSCS provides a solid evidence-based explanation of the various factors involved in a successful rotator cuff and/or subacromial impingement program (“Glenohumeral Joint Biomechanics and Rehabilitation Implementation”). A comprehensive impairment-based program focused on muscular strengthening, neuromuscular control, endurance, joint hypo/hypermobility, and surgical precautions must be implemented to treat pathologies related to the glenohumeral joint.