|Home | About | Journals | Submit | Contact Us | Français|
Codman's pendulum exercises are commonly prescribed after shoulder surgery and injury to provide grade I and II distraction and oscillation resulting in decreased pain, increased flow of nutrients into the joint space, and early joint mobilization. Many shoulder protocols suggest that weight may be added to these pendulum exercises as rehabilitation progresses, however, very few guidelines exist to stipulate how much weight should be added.
To determine if added weight affected the subject's ability to relax the shoulder musculature during pendulum exercises.
Twenty-six participants, ages 20 to 56 years old (mean 32.26, ± 8.51 years) were divided into two groups, nine pathological and 17 non-pathological. The muscle activity (EMG) of four variations of Codman's pendulum exercises 1) wrist suspended 1.5 kg weighted-ball, 2) hand-held 1.5 kg dumbbell, 3) hand-held 1.5 kg weighted-ball, and 4) no weight were recorded in each muscle.
When grouped across all patients and all other factors included in the ANOVA, the type of pendulum exercise did not have a significant effect on shoulder EMG activity regardless of patient population or muscle tested. Generally, the supraspinatus/upper trapezius muscle activity was significantly higher than the deltoid and infraspinatus activity - especially in the patients with pathological shoulders
Performing the exercises with added weight did not result in significant increased shoulder EMG activity for the deltoid and infraspinatus muscles in subjects with and without shoulder pathology. However, patients with shoulder pathology had greater difficulty relaxing their supaspinatus/upper trapezius muscle group during Codman's pendulum exercises than healthy subjects.
Early joint mobilization plays an important role in the rehabilitation of the injured shoulder for the return of normal kinematics and shoulder function.1 Prolonged immobilization may predispose patients to muscle atrophy and poor neuromuscular control. The use of early joint motion can help prevent adhesions and contractures, especially pertaining to the periarticular connective tissue.1–3 Passive range of motion is typically prescribed during postoperative care and early rehabilitation of the injured shoulder to initiate early joint motion.3 The goal of an early rehabilitation protocol is to provide motion at the glenohumeral joint, while maintaining relative inactivity of the repaired or injured muscles and tendons. These goals also aim to minimize excessive tension at the suture line, or site of injury, and prevent adverse effects on already injured musculature.1–6
Codman's pendulum exercises are commonly prescribed after shoulder pathology to provide grade I and II distraction and oscillation resulting in decreased pain, increased flow of nutrients into the joint space, and early joint mobilization.5,6 Shoulder protocols, as well as Codman5 himself, suggest that weight may be added to these pendulum exercises as rehabilitation progresses, however, very few guidelines exist to stipulate when, and how much weight should be added.6
The use of a continuous passive motion device (CPM) is not always available or appropriate after a shoulder injury or surgery, but the replication of the motion without the activation of the injured or sutured muscles is crucial for safe rehabilitation of the injured or post surgical shoulder. Typically, patients after shoulder injury or surgery have a difficult time relaxing the shoulder musculature and performing Codman's pendulum exercises correctly. Dockery et al2 evaluated the different rehabilitation protocols using electromyographic (EMG) analysis of the rotator cuff muscles to determine if different protocols promoted passive motion and how much rotator cuff activation occurred. Using a sample size of ten healthy volunteers, the authors tested various exercises commonly used postoperatively following shoulder surgery using surface electrodes. They concluded that therapist assisted exercises and Codman's pendulum exercises showed activity that was not significantly different from that of a CPM machine. The authors further concluded that these exercises have similar safety as CPM for obtaining early passive range of motion without disrupting the rotator cuff.
In a similar study, McCann et al1 used intramuscular fine wire EMG electrodes to examine shoulder muscles activity during passive, active, and resistive rehabilitation exercises. Passive exercises included Codman's pendulum exercises and EMG activity was categorized as: 1. minimal - if less than 20% of the muscle activity needed to raise a 2.25kg weight into abduction was elicited, 2. moderate - if 20%-50% of abduction activity was elicited, and 3. maximal - if greater than 50% was elicited. The results indicated that only the Codman's pendulum exercise demonstrated minimal muscle activity.
Very few studies have been conducted regarding this frequently used shoulder exercise during the early stages of rehabilitation. The purpose of this study was to determine if shoulder muscle activity increased with the addition of weight to Codman's pendulum exercises and if a difference was present in the muscle activation between the pathological shoulder and the healthy shoulder. Additionally, this study examined various methods of performing Codman's exercises and the effect on the activity of the shoulder musculature.
Twenty six participants (14 females, 12 males), ages 20 ± to 56 years old (mean age 32.3 ± 8.5 years) were recruited. Upon review and approval by the Lenox Hill Hospital Institutional Review Board, informed consent was obtained from each participant. Interested volunteers were asked a series of questions in order to categorize the participants into either the pathological or non-pathological grouping. Nine participants who had pathological shoulders (mean age 39.3 ±12.5) and seventeen participants with non-pathological shoulders (mean age 30.4 ± 6.2) made up the two groups. Participants in the pathological group were classified as having a shoulder dysfunction and/or pain within the first 6 weeks of onset or were in their first 6 weeks postoperatively. The group of pathological shoulders consisted of two patients with superior labral anterior posterior repairs (SLAP), two with shoulder dislocations, one acromial clavicular joint (AC) decompression, one corococlavicular ligament reconstruction, one shoulder impingement, one post surgical capsular shift and acromioplasty, and one post-surgical thermal shrinkage. Participants in the non-pathological group were classified as anyone who did not have a shoulder dysfunction, surgery, or pain within the past year.
Surface electrodes were positioned on the supraspinatus/upper trapezius, infraspinatus, and middle deltoid following standard surface electrode placement.7 The EMG sites were prepared according to standard protocol using a razor to remove hair, an alcohol pad to clean the skin, and an abrasive pad to abrade the skin.8 Two surface electrodes were placed 2.54cm (1 inch) apart on each of the targeted muscles to prevent the electrodes from touching with motion. The placement of the electrodes was determined by referencing the work of the Delagi and Perotto.7 The infraspinatus electrode placement was the midway point of the spine of the scapula and measuring two finger widths posteriorly and inferiorly from the middle of the spine. The supraspinatus/upper trapezius electrode placement was determined by palpation of the middle portion of the spine of the scapula, then moving superiorly two finger widths in the supraspinatus fossa. The middle deltoid placement was halfway between the acromion and the deltoid tubercle. A ground electrode was placed on the olecranon process of the arm being tested. Muscle activity output was checked prior the initiation of testing by having the participant contract each muscle to ensure correct electrode placement.
The EMG signals were band pass filtered from 10 to 500 Hz and sampled at 1000 Hz, with a common-mode rejection ratio of 130 dB (Telemyo, Noraxon, Scottsdale, AZ). A maximum voluntary isometric contraction (MVIC) for each muscle was performed for normalizing muscle activity during the Codman exercises.9,10 The MVIC was taken with the patient positioned in the standard manual muscle position for the best isolation of the targeted muscle.11 Shoulder abduction for the deltoid and supraspinatus/upper trapezius was recorded with the arm starting at the participants side, standing 30° away from a pillow held on the wall. The MVIC was recorded as the participant abducted the arm and pressed as hard as they could into a pillow against the wall. External rotation for the infraspinatus was recorded with the subject standing near a pillow held against a wall with elbow flexed to 90° and shoulder held in 0° abduction and maximally pushing the wrist outward into the wall externally rotating the arm. Participants in the pathological group were asked to perform the MVIC on the uninvolved shoulder to avoid complications with the increased muscle activity of the involved shoulder musculature, while participants in the non-pathological group were asked to perform the MVIC on the shoulder being tested.12
Participants in both the pathological and non-pathological groups were instructed to perform four variations of Codman's pendulum exercises; Suspended ball weight of 1.5 kg (3.3 lbs) (Figure 1), hand-held dumbbell of 1.5 kg (Figure 2), hand-held ball weight of 1.5 kg (Figure 3), and no weight (Figure 4). The participant was randomly assigned an order of exercises.
Participants were positioned with the non-tested arm resting on a table and the upper extremity to be tested hanging down for free movement. Trunk flexion at the hips was kept at a 75-degree angle from the upright vertical position as measured using a standard goniometer. The degree to which the trunk was flexed at the hips was modified from the traditional protocol of 90 degrees to allow for ground clearance of the suspended weight as it dangled from the arm. The “pendulum” or swinging motion was initiated by having the participant move their trunk slightly back and forth until motions of internal circumduction and then external circumduction were achieved. A circle was placed on the floor for guidance to control for the amount of circumduction that the participant would achieve during the testing. Shoulder range of motion limits were set to the minimum ability of the most involved shoulder pathology to avoid any possible complications. The speed of the arm swing was controlled for each participant using a loud beat on an electric metronome set at 40 beats per minute. Three trials for each parameter were performed to allow participants to become comfortable with the motion and the testing procedure. The third trial was used for data collection. Each participant performed five clockwise and five counter clockwise circles for each of the four parameters tested.
The EMG signals were acquired using the Noraxon TeleMyo telemetered EMG system (Noraxon USA, Scottsdale, AZ). The signals were low pass filtered at 500 Hz and high pass filtered at 10Hz. The EMG readings were sampled at 1kHz and analyzed using Noraxon Myosoft software. A 100ms moving average RMS function was applied to the raw EMG signal. The average RMS for the duration of the trial was computed by integrating the RMS and dividing the area by the time, thereby, producing the average amplitude. This process was conducted for all activity during the last trial and expressed as a percent of that particular muscle's MVIC. Internal and external circumduction were calculated for the non-pathological group since both arms were tested (i.e., left shoulder internal circumduction compared to right arm external circumduction). Internal and external circumduction were also calculated for each arm individually for the non-pathological group to verify the related muscle firing pattern was not different. The pathological group did not require these calculations since only the contralateral arm was tested.
The effect of the different types of Codman exercises on muscle activity was examined using a 4×3×2 (exercise type × shoulder muscle group × pathology group) mixed model analysis of variance (ANOVA) with Bonferroni corrections for pairwise comparisons. Separate exercise type vs pathological group ANOVAs were performed on each muscle separately to further examine the data. Greenhouse-Geisser corrections were applied to significant ANOVAs that did not meet Mauchly's sphericity assumption in order to reduce the likelihood of a type I error. An alpha of 0.05 was set a priori.
When grouped across all patients and all other factors included in the ANOVA, the type of pendulum exercise did not have a significant effect on shoulder EMG activity regardless of patient population or muscle tested (effect of exercise type - p=0.79, exercise type by shoulder muscle interaction - p=0.72). Generally, supraspinatus/upper trapezius muscle activity (17% MVIC) was significantly higher than deltoid (6%) and infraspinatus activity (7%) (muscle effect - p=0.001).
When each muscle was analyzed separately for the effect of added weight distraction during Codman's pendulum exercises (comparing the three weighted conditions to the non-weighted condition), no significant increase in shoulder muscle activity occurred for any of the weighted conditions in the infraspinatus (Figure 5; p=0.39), supraspinatus/upper trapezius (Figure 6; p=0.36) and two of the exercise conditions for the deltoid. However, deltoid activity was increased during the pendulum exercise with a dumbbell (Figure 7; p=0.04).
Further analysis indicated that the pathological group had significantly greater muscle activity in the infraspinatus (Figure 5; p=0.041) and supraspinatus/upper trapezius (Figure 6; p=0.03) compared with the non-pathology group. No such significant difference between the pathological and non-pathological groups occurred in the deltoid group (Figure 7; p=0.11).
Passive shoulder motion is regarded as standard early rehabilitation in patients postoperatively, as well as patients not undergoing surgery. Early joint mobilization plays an important role in the rehabilitation of the patient with an injured shoulder for the return of normal kinematics and shoulder function.1,2 Dockery et al2 investigated the different shoulder rehabilitation protocols using surface EMG analysis of the rotator cuff muscles to determine how much rotator cuff activity the protocols promoted. The authors concluded that shoulder muscle activity during Codman's pendulum exercises was not significantly different from that of a continuous passive motion (CPM) machine. The study conducted by Dockery et al2 used similar methodology (including the use of surface EMG) and found similar results as this present study.
In a similar study, McCann et al1 used intramuscular fine wire EMG electrodes to examine shoulder muscle activity during passive, active, and resistive rehabilitation exercises. Only the pendulum exercise consistently showed minimal shoulder muscle activity defined as eliciting less than 20% of the muscle activity needed to raise a 2.25kg weight in abduction. The results of the present study are in agreement with the study of McCann et al,1 indicating that the type of pendulum exercise minimally effects shoulder muscle activity. A small increase in deltoid activity occurred in the group with shoulder pathology compared to the group without pathology during the pendulum exercise performed with the dumbbell but the infraspinatus and supraspinatus/upper trapezius did not show a similar effect. In comparing the pendulum exercises within each muscle with no weight to the three pendulum exercises performed with weight it was apparent that the added distraction force by the addition of weight did not increase shoulder muscle activity. All muscle activity - for each type of exercise and for pathological and non-pathological shoulders exercise - was less than 20% MVIC except for the supraspinatus/upper trapezius in the pathological group. Therefore, the majority of the Codman's pendulum exercise for early shoulder rehabilitation fell below the minimal category established by McCann et al.1
The other clinically relevant finding was that the supraspinatus/upper trapezius and infraspinatus muscle activity were significantly higher in the group with shoulder pathology compared to the group without pathology. However, although a significant difference was found between the group with pathology and the group without pathology for the infraspinatus muscles, the percent of muscle activity was below 15% for all types of exercise in both groups. Therefore, the infraspinatus muscle appears to be able to relax during pendulum exercises.
The only muscle group that appeared to be unable to relax during the pendulum exercises was the supraspinatus/upper trapezius. The magnitude of muscle activity in the group with shoulder pathology indicated that these subjects were not able to sufficiently relax their supraspinatus/upper trapezius muscle. Mean supraspinatus/upper trapezius activity was 39% MVIC for the pendulum exercise performed without a weight, 20% when performed with a suspended weight, 25% when performed with a grip weight, and 30% when performed with a dumbbell. Corresponding values in the group without pathology were 10%, 13%, 12%, and 11% (Table 1). This response may reflect a guarding response in patients with pathology. Biofeedback training may be necessary to facilitate relaxation in the supraspinatus/upper trapezius muscles during the pendulum exercises.
Since the results of this present study were largely negative with respect to the effect added weight on muscle relaxation it is important to assess whether this effect might have been subject to a type II error. Therefore, post hoc analyses were performed to determine the magnitude of difference in EMG activity that could be detected as significant (p<0.05) with 80% power. Based on the standard deviation of the difference in EMG activity (% MVIC) between the different types of pendulum exercises these detection thresholds were 5% for the supraspinatus/upper trapezius and 2.5% for the infraspinatus and deltoid. These analyses indicate that there was sufficient power to detect clinically relevant differences in EMG activity between the different types of pendulum exercises.
Limitations of the study include use of surface EMG to measure muscle activity, especially in the supraspinatus muscle. To say that the muscle activity of the supraspinatus was isolated without interference from the upper trapezius and possibly the posterior deltoid is problematic. Therefore, the category of supraspinatus/upper trapezius was used. Despite this limitation, we believe that this study still provides valuable information related to muscle activity during the performance of Codman's pendulum exercises when adding weight. An additional limitation was the sample size. A larger sample size, especially in the group with pathological shoulders, would have been nice to compare to the control group.
The results of this present study indicate that adding a 1.5kg weight had no significant impact on shoulder muscle activity (as demonstrated by EMG analysis) during pendulum exercises for the deltoid and infraspinatus muscles for subjects with or without shoulder pathology. However, the supraspinatus/upper trapezius muscle group was clearly not relaxed in patients with shoulder pathology during any of the pendulum exercise.