Regaining motion among patients with shoulder pathology - are all exercises equal?

Abstract

Background

Little information exists to guide the choice of exercise for regaining shoulder range of motion (ROM). The purpose of this study was to compare the maximal ROM reached, pain and difficulty associated with 4 commonly prescribed exercises.

Methods

Forty (9 females) patients with various shoulder disorders and a limited flexion ROM performed 4 exercises for regaining shoulder flexion ROM in a randomized order. Exercises included the self-assisted flexion, forward bow, table slide and rope-and-pulley. Participants were videotaped while performing all exercises and the maximal flexion angle reached during each exercise was recorded using Kinovea motion analysis freeware (Kinovea 0.8.15). Pain intensity and the perceived level of difficulty associated with each exercise were also recorded.

Results

The forward bow and table slide generated significantly greater ROM compared with the self-assisted flexion and rope-and-pulley (P ≤ 0.005). The self-assisted flexion was associated with a greater pain intensity compared with the table slide and rope-and-pulley (P = 0.002) and a greater perceived level of difficulty compared with the table slide (P = 0.006).

Conclusions

Due to the greater ROM allowed, and similar or even lower level of pain or difficulty, clinicians may wish to initially recommend the forward bow and table slide for regaining shoulder flexion ROM.

Introduction

Management of shoulder pathologies often involves several weeks of complete or relative immobilization followed by a progression of exercises initially focused on regaining shoulder range of motion (ROM). Given its predominant role in most basic and instrumental activities of daily living,^1–3 shoulder flexion ROM is emphasized early during the rehabilitation process.^4–6 Existing protocols advocate a variety of exercises in order to regain shoulder flexion ROM, however, very little information exists to guide the selection of any particular exercise.

Much of the literature concerning exercises used to regain shoulder flexion ROM has focused on the associated level and pattern of muscle activation.^7–15 While such studies help classify the mode of the exercise (i.e. passive, active-assisted or active) and thus estimate the associated tissue loading, they do not provide information on the very purpose for which these exercises are performed, namely, their potential in fully moving the shoulder through its available ROM. Existing exercises differ greatly in terms of body position (e.g. supine-lying, sitting or standing), the type of assistive device used (e.g. opposite hand, a table, wand or pulley handle), as well as the moving segment (e.g. humerus versus trunk). Given this variability it is conceivable that different exercises may allow differing degrees of shoulder motion, as well as be associated with differing levels of pain or difficulty. An understanding of these potential differences may assist clinicians in selecting exercises to promote greater joint excursion while minimizing pain intensity and level of difficulty.

Given that glenohumeral motion often becomes limited and substituted by greater scapulothoracic motion among patients with various shoulder pathologies,^16–18 it seems particularly important to identify exercises with potential to promote greater glenohumeral excursion. Therefore, the primary goal of this study was to compare the maximal ROM reached during 4 commonly prescribed exercises for regaining shoulder flexion among patients with various shoulder pathologies. Secondary goals were to compare the intensity of pain and the level of difficulty associated with the performance of each of these exercises.

Materials and methods

Participants

Consecutive patients visiting the outpatient clinic of a shoulder surgery unit within a large metropolitan centre were recruited for this study. Inclusion criteria included age ≥ 18 years, unilateral shoulder pathology and a clear deficit in passive shoulder flexion ROM as compared with the uninvolved side. Patients were excluded if passive shoulder flexion was greater than 160°, if they could not assume a standing position (e.g. bound to a wheelchair), if ROM was contra-indicated, or if they could not use their uninvolved upper extremity to perform the investigated exercises. Sample size calculation was performed using GPower version 3.1 (Heinrich Heine Universitat, Dusseldorf). Based on the intent to detect a moderate effect size (≥0.25) between the different ROM exercises using a 2-tailed test, a P-value ≤ 0.05 and a desired power (β) of 95%, the required sample size was estimated to be 36 participants.

All participants received verbal and written explanation regarding the purpose and procedures of the study before signing an informed consent form approved by an institutional review board.

Examiner

All procedures were carried out by a single physical therapist with over 20 years of experience in the assessment and management of patients with shoulder disorders.

Procedure

Participants first provided demographic and history information pertaining to their shoulder disorder. Each participant then continued to perform 4 exercises commonly prescribed to increase shoulder flexion ROM: 1.) self-assisted flexion; 2.) table slide; 3.) forward bow; and 4.) rope-and-pulley. The exercises were performed in a random order based on a pre-established random list of numbers generated from www.random.org. Randomization was performed to minimize the effects of pain, fatigue, and/or tissue conditioning. During each exercise participants were asked to reach their maximum tolerable range. Following a demonstration of the exercise participants were given 3 familiarization trials to ensure proper performance of the exercise followed by 3 additional repetitions during which they were asked to gradually reach their maximum tolerable ROM. The exercises were performed as follows:

• Self-assisted flexion – The participant assumed a supine lying position with both hands resting over the chest and the fingers interlocked. The participant then used the uninvolved hand to pull the involved shoulder into maximum tolerable flexion (Figure 1).

• Table slide – The participant was seated on a standard (45 cm) chair with the forearms resting over a table surface 80 cm high and the fingers of both hands interlocked. The participant leaned his/her trunk forward sliding both forearms over the table until reaching his/her maximum tolerable flexion (Figure 2).

• Forward bow – While facing a table (80 cm high), each participant placed both hands on the table surface with the elbows straight and the forearms pronated. The participant then stepped back, leaned backward and lowered the chest toward the floor until reaching his/her maximum tolerable flexion (Figure 3).

• Rope-and-pulley – The participant sat with his/her back to a pulley system anchored over the top of a door. The initial position was with the involved shoulder adducted by the side, the elbow bent and the hand grasping the pulley handle at chest level. The uninvolved arm was elevated so that the opposite handle was grasped overhead. The participant then extended the uninvolved shoulder thus pulling the involved shoulder into maximum tolerable flexion (Figure 4).

Figure 1.

Self-assisted flexion.

Figure 2.

Table slide.

Figure 3.

Forward bow.

Figure 4.

Rope and pulley.

Following the execution of each exercise participants rated the intensity of pain experienced during the exercise on an 11-point (0–10) numeric pain rating scale,^19,20 as well as the perceived level of difficulty in performing the exercise using a 5-point (0–4) scale (0 = no difficulty; 1 = mild difficulty; 2 = moderate difficulty; 3 = severe difficulty; 4 = inability to perform).

Flexion ROM measurement setup: a hand-held camcorder (GC-PX100, JVC, Yokohama, Japan) was used to capture a 2-s video from a side view while participants held the maximal flexion ROM during each exercise. Shoulder flexion was measured as the angle between the lateral aspect of humerus and the lateral aspect of the scapula rather than the trunk. This was performed in order to minimize trunk and/or scapula substitutions that would artificially inflate the ROM. Three round paper stickers (20 mm diameter) were used to facilitate the ROM measurement. Prior to performing the exercises one sticker was placed over the lateral epicondyle of the humerus while another sticker was placed over the estimated centre of the glenohumeral joint, defined as the midpoint between the posterior aspect of the acromion and the axillary fold while the shoulder was flexed 90°. ²¹ Once participants reached their maximal flexion angle during each exercise the examiner palpated and marked the inferior angle of the scapula with a third paper sticker. The examiner then immediately proceeded to capture a 2-s video from a side-view capturing all 3 landmarks in the maximal ROM position. In order to achieve a consistent angle of view, the examiner first placed the camera flush against the sticker overlying the centre of the glenohumeral joint. The examiner then stepped back 3 feet while maintaining the camcorder level and at a 90° to the paper sticker before recording the video capturing all 3 landmarks. The entire procedure did not take longer than 10 s from the moment each participant reached his/her maximal ROM during each exercise.

ROM measurement: the maximal flexion ROM reached during each exercise was measured with a freeware motion analysis software (Kinovea 0.8.15). The angle measurement feature of the software was used as follows: first, the point of the angle was placed over the sticker overlying the centre of the glenohumeral joint. One arm of the angle was then extended along the lateral border of the scapula bisecting the sticker overlying the inferior angle of the scapula, while the second arm of the angle was extended over the lateral aspect of the arm bisecting the sticker overlying the lateral humeral epicondyle (Figure 1–4). In order to determine intra-rater reliability the same examiner repeated the measurement on the video recordings of the first 10 participants. The second measurement was performed 1 month after the first measurement in order to minimize recall bias. Intraclass correlation coefficients (ICC) with 95% confidence interval (CI) and minimal detectable change at 95% confidence (MDC⁹⁵) were 0.99 (0.99–1.00) and 2.8° for the self-assisted flexion, 0.99 (0.99–1.00) and 2.3° for the table slide, 0.99 (0.99–1.00) and 2.1° for the forward bow, and 0.99 (0.99–1.00) and 2.1° for the rope-and-pulley, respectively.

In order to validate this measurement procedure, a pilot session was undertaken in which the video-based measurement was compared with a 3-dimensional (3D) motion capture measurement. Briefly, 2 retro-reflective markers (14 mm in diameter) were placed using double sided adhesive tape over the lateral epicondyle of the humerus and the estimated centre of the glenohumeral joint (i.e. midpoint between the posterior aspect of the acromion and the axillary fold in 90° of shoulder flexion) of a healthy individual. This individual assumed 8 self-selected shoulder flexion angles during each of the 4 ROM exercises. Once each self-selected position was assumed the inferior angle of the scapula was palpated by the examiner and marked with a third retro-reflective marker. Each self-selected elevation angle was then measured with the video-based procedure used in the primary analysis as well as with a Qualysis Motion Capture System (Qualysis Medical AB, Gothenburg, Sweden) equipped with 6 Pro-Reflex MCU 1000 cameras with a capturing rate of 120 frames per second. ICC with 95% CI were calculated to determine agreement (convergent validity) between the video-based and the 3D motion capture measurement of the sagittal-plane angle formed by the 3 retro-reflective markers. The ICC (95% CI) for the self-assisted flexion was 0.98 (0.86–1.00), for the table slide 0.84 (0.39–0.97), for the forward bow 0.99 (0.98–1.00), and for the rope and pulley 0.98 (0.88–1.00), representing excellent agreement between the 2 measurements.

Statistical analysis

Descriptive statistics were used to summarize all data with measures of central tendency and dispersion for interval variables and frequency counts for categorical variables. Non-parametric analyses were used due to a non-normal distribution of some of the variables based on Shapiro-Wilk tests.

For the primary aim of the study a Friedman test was used to assess for differences in the ROM achieved during the 4 ROM exercises. In case of a statistically significant analysis separate pair-wise Wilcoxon tests were used to assess for differences between the exercises using a Bonferroni-corrected P-value of 0.008 (0.05/6). Similar analyses were performed to assess for differences in pain intensity and perceived level of difficulty with each exercise.

All analyses were performed using SPSS version 21 (SPSS, Inc, Chicago, IL) with an a-priori level of significance of P ≤ 0.05.

Results

Forty patients (9 females) were recruited for the study. The average ± SD age, height, and weight of participants were 47.9 ± 16.1 years, 174.4 ± 8.0 cm, and 78.6 ± 14.0 kg, respectively. Twenty-eight participants were seen as part of their post-operative follow-up (9 following open-reduction internal fixation of a proximal humerus fracture, 9 following a shoulder stabilization procedure, 8 following rotator cuff repair, 1 following open-reduction internal fixation of a glenoid fracture, and 1 following reverse total shoulder arthroplasty due to fracture sequalae); 6 participants were seen for follow-up of a conservatively managed proximal humerus fracture, 3 participants were diagnosed with adhesive capsulitis, 2 participants were seen for follow-up of a conservatively managed glenoid fracture, and 1 participant was seen for a rotator cuff tear. Three participants were unable to perform the self-assisted shoulder flexion exercise due to excessive pain while 2 other participants were unable to perform the rope-and-pulley exercise also because of excessive pain.

The ROM achieved with each exercise, as well as the intensity of pain and perceived level of difficulty with each exercise, are summarized in Table I. The analysis regarding ROM was statistically significant (P < 0.001) indicating differences in the maximum ROM achieved during the exercises. Post-hoc pairwise comparisons indicated that the forward bow allowed a greater flexion ROM compared with the self-assisted flexion (P < 0.001) and the rope-and-pulley (P < 0.001). In addition, the table slide allowed a greater flexion ROM compared with the self-assisted flexion (P = 0.005) and the rope-and-pulley (P < 0.001). No differences were found between the self-assisted flexion and the rope-and-pulley (P = 0.87) or between the forward bow and the table slide (P = 0.01).

Table 1.

Mean ± SD shoulder flexion range-of-motion, pain intensity and perceived difficulty during each exercise.

Exercise	Range-of-motion, °	Pain intensity, (0 – 10)	Perceived difficulty, (0 – 4)
Forward bow a	123.1 ± 16.0	5.1 ± 2.7	1.3 ± 0.9
Table slide b	121.2 ± 13.3	4.4 ± 3.1	1.0 ± 0.9
Rope-and-pulley	116.9 ± 16.2	5.0 ± 2.9	1.5 ± 1.1
Self-assisted flexion c, d	116.4 ± 18.4	5.9 ± 2.6	1.6 ± 1.2

^a Greater ROM than self-assisted flexion and rope-and-pulley (P < 0.001).

^b Greater ROM than self-assisted flexion (P = 0.005) and rope-and-pulley (P < 0.001).

^c Greater pain intensity than table slide and rope-and-pulley (P = 0.002).

^d Greater perceived difficulty than table slide (P = 0.006).

The analysis regarding the level of pain associated with each exercise was statistically significant (P = 0.001) indicating differences between exercises. Post-hoc pairwise comparisons indicated the self-assisted flexion was associated with greater pain intensity compared with the table slide (P = 0.002) and the rope-and-pulley (P = 0.002). No other statistically significant differences were found between any of the other exercises.

The analysis regarding the level of perceived difficulty with each exercise was statistically significant (P = 0.04). Post-hoc pairwise comparisons indicated that the self-assisted flexion was associated with a greater degree of perceived difficulty compared with the table slide (P = 0.006). No other comparison reached a statistically significant difference.

Discussion

Commonly prescribed exercises for regaining shoulder flexion ROM differ in the maximal ROM reached, as well as in the associated intensity of pain and perceived level of difficulty. Closed-chain exercises such as the forward bow and table slide, in which the arm is supported and ROM is introduced via movement of the trunk, allow patients to stretch further compared with open-chain exercises such as the self-assisted flexion and rope-and-pulley in which the involved arm is moved upon a stationary trunk. The differences in ROM between the closed- and open-chain exercises exceed the MDC⁹⁵ associated with their respective measurements suggesting these differences are unlikely to represent measurement error. In addition, the self-assisted flexion is associated with a greater pain intensity compared with the table slide and rope-and-pulley, as well as a greater perceived level of difficulty compared with the table slide.

As the self-assisted flexion and rope-and-pulley involve movement of the injured arm on a stationary trunk, greater shoulder muscular activity is expected to be involved. Accordingly, the rope-and-pulley has been found to elicit greater rotator cuff and deltoid activation compared with the forward bow or table slide,^11,21 and considerable activation of the latissimus dorsi has been documented during the self-assisted shoulder flexion, most likely to control arm movement past 90° of shoulder flexion. ¹³ By contrast, as the forward bow and table slide rely on trunk movement to mobilize the shoulder, these exercises may help promote a greater sense of security and increased willingness to move. Furthermore, during the open-chain exercises neither segment of the glenohumeral joint is truly stabilized and as glenohumeral motion is exhausted compensatory trunk extension is likely to occur. ²² By contrast, as during closed-chain exercises the arm is directly (table slide) or indirectly (forward bow) stabilized via contact with a support surface, the forward movement of the trunk may lead to greater separation between the scapula and the humerus resulting in greater joint excursion. Interestingly, a previous study has also found the forward bow allowed patients following superior labral repair to reach the greatest flexion ROM compared with several other exercises (including the rope-and-pulley). ²¹

Consensus statements regarding post-operative rehabilitation after shoulder replacement and rotator cuff repair recommend all or most of the exercises included in the current investigation with no particular preference for any specific exercise.^6,23 Several post-operative and non-operative shoulder rehabilitation protocols also recommend one or another of the exercises included in this study.^24–26 These recommendations are based on shoulder muscle activation recorded during these exercises or on the personal preference of the authors. The findings of this study add a new perspective on the possible utility of these exercises in regaining shoulder ROM. It seems exercises performed under closed-chain conditions allow patients to stretch further into their ROM and may therefore be preferable to open-chain ROM exercises. The greater ROM reached through closed-chain exercises combined with a similar, and sometimes lower intensity of pain or level of difficulty may also promote greater patient adherence which is likely to result in greater ROM gains as well.

Our study has several limitations. First, as we measured the ROM reached during the performance of all exercises, our findings do not provide evidence for the relative efficacy of these exercises in regaining shoulder ROM over the course of a rehabilitation programme. More specifically, it is unclear whether differences in the ROM reached during the closed- versus the open-chain exercises will eventually translate into greater or earlier gains in flexion ROM over time. A prospective comparison between these exercises would be a valuable line of future research. Second, although we attempted to exclude trunk and/or scapular substitution by measuring the angle between the lateral aspect of the scapula and the humerus, it is clear our measurements exceed the ROM expected to occur within the glenohumeral joint in isolation. Third, as our data was collected from a sample of patients affected by different conditions it is possible that a more homogenous group of patients would have displayed different findings. Nevertheless, ROM exercises are prescribed based on the presence or absence of a ROM limitation rather than based on a specific underlying pathology. As a result, recruiting patients with different disorders seemed justified so long as a ROM limitation was present. Finally, the scale used for measuring the perceived level of difficulty associated with each exercise has not been previously validated and caution should be used when interpreting related findings.

Conclusion

ROM exercises performed under closed-chain conditions allow patients to reach greater flexion ROM without incurring extra pain or difficulty. Clinicians may wish to prescribe closed-chain ROM exercises such as the forward bow or table slide first to patients who need to regain shoulder flexion ROM.