The immediate effect of muscle release intervention on muscle activity and shoulder kinematics in patients with frozen shoulder: a cross-sectional, exploratory study

This is an exploratory and cross-sectional study. Twenty patients suffering from unilateral FS and 20 asymptomatic subjects were recruited from the Taipei area, Taiwan. Based on data from previous studies [11], the sample size of 20 participants in each group was considered adequate to detect between group and within group difference with a power of 80% (significance level = 0.05). Inclusion criteria for the patient group included [1] medical diagnosis of FS (by physicians specializing in Orthopedics or Physical Medicine), [2] pain and stiffness over the affected shoulder region for more than three months, [3] no resting pain or night pain in the affected shoulder region, and willingness to participate in this study. Asymptomatic subjects were recruited with age, gender, BMI (body mass index), and tested shoulder matched with the FS group. Exclusive criteria for all subjects were: [1] history of surgery or fracture of the shoulder complex, [2] shoulder joint dislocation, [3] rheumatoid arthritis, [4] osteoarthritis of the particular shoulder joint, [5] cervical radiculopathy, or [6] shoulder ROM (range of motion) limitation due to stroke or spinal cord injury.

Ethical clearance was obtained from Taipei City Hospital Institutional Review Board (IRB number TCHIRB-1030708-E). The testing procedures were fully explained, and written informed consent was obtained for all participants before commencement of the study. This experiment was conducted at the Musculoskeletal and Sports Sciences Laboratory, National Yang-Ming University, Taiwan.

We used an 8 channel FM/FM Telemetric EMG system (Telemyo 2400, Noraxane USA) to record muscle activation. The input impedance of the system was 10 MΩ, common mode rejection ration of 85 dB and gain of 2000. All signals were converted by an analog-to digital (A/D) converter (NI PCM-CIA 6036 E, USA; 12-bit resolution) and the sampling rate was 1500 Hz. The raw electromyography (EMG) signals were collected by silver/silver chloride pre-gelled surface electrodes (Blue Sensor P-00-S, Ambu Inc., USA) with 2 cm inter-electrode distance. The electrodes were placed at the midway between the spinous process of 7th cervical vertebra, the center of the acromion, and the spinous process of eighth thoracic spine for the lower trapezius (LT); at the center of the infraspinous fossa for the infraspinatus muscle (ISp); at the middle of the muscle belly along the lateral border of the scapula for the teres major (TM) [17], and 3.5 cm medial to the anterior axillary line for the pectoralis major (PM) [18]. The ground electrode was attached to the head of the clavicle.

The Liberty electromagnetic tracking system (Polhemus, Colchester, USA) was used to collect the three-dimensional shoulder kinematics at 120 Hz. Three sensors were used and attached to the spinous process of the 7th cervical vertebra, posterior-lateral acromion, and lower 1/3 aspect of the humerus with adhesive tape [19]. A pen-like stylus was used to digitize the palpated bony landmarks to define the anatomical coordinate systems based on the suggestions of International Society of Biomechanics [19].

The FS group received the electrical heating pad for 15 min at a temperature between 42 and 45 degrees Celsius, followed by one-session of manual muscle release (PM, UT, ISp, TM, and posterior deltoid) for about 30 min right after the initial examination. The target muscle of intervention was positioned in the lengthened position, while the physical therapist used the elbow and fingers to give sustained pressure directly on the most tender points at the muscle belly for 60 to 90 s until the physical therapist felt the target muscles start to release or the patient felt the pain decrease (Fig. 1) [15, 16]. The same licensed physical therapist (Liao PW), who has had more than two years practicing this technique performed all the treatments.

Figure 2 shows the flow of the study. All subjects performed a 10 min warm-up with a hand cycle. Subjects were free to choose their comfortable speed and resistance level. This was followed by measurement of surface EMG (PM, IPs, TM, UT, and LT) and shoulder kinematics (humeral elevation, and scapular tilting and upward rotation) during three functional tasks in the sitting position with the trunk well stabilized by the chair, belt, towel, and foam to avoid any compensatory trunk motion. The functional tasks were: scaption (arm lift in the scapular plane as far as they could), hand to neck (touching the back of the neck using the palm), and thumb to waist (touching the thumb to the back of the waist at the level of the 12th thoracic spine). Before data recording, subjects were asked to practice three to five times to ensure no compensation occurred during the task. Each task was performed at subjects’ comfortable speed three times with 30 s rest in between.

Afterwards, the subject was asked to hold a one-kilogram weight at 80° scaption for five seconds three times while the EMG data of the UT, LT, and ISp during these reference voluntary contractions (RVCs) were recorded [14]. We used a 1 Kg weight for the RVCs because our participants with FS found any weight more than 1 Kg too painful or too heavy to hold, and the RVC normalization method was able to produce reliable EMG data for UT, LT, and ISp without over-fatiguing the participants with shoulder pain [14]. The EMG data during three repetitions of maximum voluntary contractions (MVCs) of the PM and TM were assessed using the standard manual muscle testing methods [17, 20]. After a three minute rest, the measurements of shoulder active and passive ROM (flexion, abduction, external and internal rotation) were performed in the supine position using a universal goniometer with an inclinometer (Table 1) [21]. The data of test-retest repeatability of the goniometer measurement including intraclass correlation coefficient (ICC(3,3)), standard error or measurement (SEM), and minimal detectable change (MDC₉₅) are summarized in Additional file 1: Table S1. Three measurements in each direction were performed and averaged for further analysis. A 10 cm visual analog scale (VAS) was used to measure the pain intensity (accuracy = 1 mm) in the FS group, with 0 = no pain and 10 = worse pain.

After the initial assessment, the locations of the electrodes and motion sensors were marked on the skin, and the electrodes and sensors were removed before intervention. The FS group then received the muscle release intervention, followed by re-attaching the electrodes and motion sensors and the post-treatment assessment of the muscle activity, shoulder kinematics, shoulder ROM, and pain. All assessment and interventions were performed by a licensed physical therapist with more than two years of clinical experience with our assessment and intervention methods.

The EMG raw data were band-pass filtered (20 to 500 Hz), rectified, and root-mean-squared (RMS) with a window of 50 ms. These data were then normalized by the RVC (UT, LT, ISp) or MVC (PM, TM) EMG data recorded at the baseline, and the mean normalized values were calculated for each task. Data of the three trials were averaged. The measurement of the surface EMG showed good test-retest reliability with ICCs [3] between 0.71 and 0.98 (Additional file 1: Table S1).

The Motion Monitor® (Innovative Sport Training, Inc., Chicago, USA) was used to analyze kinematic data. The Euler angles of the rotational matrices of the humerus and scapula related to the thorax were then calculated. The humeral movement was described firstly about the Y-axis of the thorax (the plane of elevation), the X-axis of the humerus (humeral elevation), and then the Y-axis of the humerus (humeral internal/external rotation). Scapular rotations along the Z-axis defined upward rotation (+), and along the X-axis defined scapular posterior tilt (+) [19]. The peak value of humeral elevation, scapular tilt and upward rotation during functional tasks were obtained and averaged for further comparison. The measurement of shoulder kinematics had a moderate to good test-retest reliability (ICC (3,3) = 0.63~0.99) (Additional file 1: Table S1).

SPSS 16.0 (SPSS Inc., Chicago, USA) was used for data analysis. The level of significance was set at .05. The multivariate analysis of variance (MANOVA) was used to compare the differences in EMG activity, shoulder kinematics, and shoulder ROM between the FS and the asymptomatic group. The repeated-measure MANOVA was used to compare all the outcome variables before and after the intervention in FS group.

Our participants with FS had restricted shoulder ROM, and the degree of limitation varied. Therefore, the target of the functional tasks was adjusted when the subjects in the FS group encountered difficulties during the tasks. We used surface EMG to quantify the level of shoulder muscle activity, and thus crosstalk from neighborhood muscles might occur and influence the results. Our outcome measures did not include the important kinematics for FS assessment, the humeral internal/external rotation, due to inadequate measurement accuracy. We did not standardize the speed of movement to avoid triggering too much pain during the assessment. This uncontrolled factor would nevertheless have an impact on the measurement of kinematics and could not be overlooked. Our data showed that the participants in the patient group were not considered severe for frozen shoulder as they averaged 96.86° ± 8.75°and 36.61° ± 13.85°for passive shoulder abduction and external rotation, which could have potentially skewed the results. We did not objectively control some parameters of the manual muscle release intervention, such as the amount of pressure applied or the total duration of the intervention. The present study only assessed the immediate effect of muscle release intervention with no control or placebo group for comparing the effects of intervention. A future study with a long-term intervention and follow up, and a randomized controlled design is needed to determine the clinical value of the muscle release intervention in patients with FS.