This was a parallel-design 2-arm, assessor-blinded randomized controlled clinical trial. Outcomes were assessed at baseline, immediately post intervention, and at week 4 post intervention. The study’s protocol was in compliance with the CONSORT guidelines.
A sample of 42 physically active adult athletes (20 males, 22 females) voluntarily participated in this study. Participants were recruited by flyers, emails, referrals, and word of mouth from a college campus and a community in Sothern California. All participants read and signed a written informed consent approved by the Institutional Review Board of Loma Linda University prior to participation. Eligible participants met the following inclusion criteria: 1) between 18 and 45 years of age; 2) performed overhead sports activities, such as volleyball, tennis, water polo, squash, baseball, swimming, or lifting as a major activity of interest; 3) participated in local and regional sport competitions 4) displayed a 10° or greater of between-shoulder difference in IR ROM (dominant versus non-dominant); and 5) with or without pain during shoulder activities. These criteria were selected based on a recent systematic review, which suggested that a lower threshold ROM deficit, based on a weighted mean GIRD of 13.8° among all injured athletes, should be utilized if the goal is to improve ROM and prevent injuries in these athletes (24). Subjects were excluded if they were: 1) still recovering from previous surgery of the shoulder and elbow complex in the past 3 months; 2) currently receiving medical intervention for the shoulder; or 3) have any critical medical condition. All experimental procedures were conducted in the orthopedic laboratory at Loma Linda University, Department of Physical Therapy from April 2017 to November 2018.
A digital inclinometer (MicroFET3, Draper, UT) was used to measure GH IR and ER ROMs. This instrument has been shown to have a high intra-rater and inter-rater reliability [Intraclass correlation coefficient (ICC) = 0.87 and ICC = 0.93, respectively] (25). Subsequently, a two-pound pressure, via an electronic push/pull dynamometer (Baseline Electronic Push/Pull Dynamometer - MicroFET3, Draper, UT), was applied by the same examiner at the end of available passive range of motion prior to taking a ROM measurement in order to standardize the shoulder internal and external rotation values. Pain was measured using the Numeric Pain Rating Scale (NPRS). NPRS has been shown to be valid, reliable, and appropriate for use in clinical practice (26).
Procedures and intervention
Forty subjects who fulfilled the inclusion criteria were recruited. After stratification, participants were randomly assigned to one of the 2 intervention groups. Group A comprised the modified sleeper stretch (MSS) group (n = 22) (Figure 2) while Group B comprised the novel stretch (NS) group (n = 20) (Figure 1). Group assignment was conducted by an independent person using a random number generator and concealed in sealed envelopes from all personnel involved in screening before randomization. Post randomization, objective measures’ assessors were blinded to group assignment. Participants were taught their assigned home-based stretching program by the investigator who explained the stretching technique to each participant. All participants were asked to demonstrate the stretch to ensure mastery. Additionally, to ensure and facilitate compliance, participants were provided a written sheet with detailed instructions and demonstrative pictures of the stretch. A log sheet was provided to monitor the number of stretches per week. Participants were also contacted weekly through phone calls and text messages as a reminder to minimize lack of compliance. All participants performed 4-weeks of stretching protocol, three times a week, with three repetitions, holding each stretch for 30 seconds, with 30 seconds rest between repetitions (27).
Figure 2. Modified Sleeper Stretch: Passive Glenohumeral Joint Internal Rotation.
Participants in Group A were instructed to perform the MSS in a side-lying position, on the side to be stretched, with the elbow flexed to 90°, shoulder elevated to 90°, and the trunk slightly rolled posteriorly 200 to 300 to place the shoulder in the scapular plane as passive IR is performed (Figure 2). Participants were instructed to allow the stretched shoulder to naturally fall into IR to the end ROM where resistance could be felt. Participants were then instructed to use the other hand to push the stretched shoulder into further IR to the point of mild stretch discomfort by applying pressure at the area of the wrist joint.
Participants in Group B were instructed to perform the NS in a supine position. Participants placed a resistance band around their knees and were instructed to open their knees. Participants were asked to bridge as high as possible, with the elbow flexed to 90° and shoulder abducted to 90° (Figure 1). The bridging maneuver shifts the body weight superiorly pinning the medial border of scapula against the thorax without directly compressing or restricting posterior shoulder structures. It is proposed that this position will allow greater freedom of motion with less discomfort. While maintaining this position, participants were asked to contract or “squeeze” their gluteal muscles and perform the stretching by actively internally rotating their shoulder to the end of ROM. Participants used the other hand to push to the point of mild stretch discomfort and simultaneously maintain contraction while progressing the stretch.
Shoulder IR was measured by positioning the participant in a side lying position, with the elbow flexed to 90° and shoulder abducted to 90°. Participants were instructed to allow the shoulder to naturally fall into internal rotation to the end ROM where resistance was felt. At this point, an electronic push/pull dynamometer was applied to the distal forearm at the end range of motion. Before recording the ROM measurement, the physical therapist applied consistent pressure (2 pounds) in an attempt to standardize and quantify the passive force to ensure reliable and valid ROM measurements. The digital inclinometer was then aligned along the ulna of the stretched arm and three trials of the internal rotation ROM measurements were taken, averaged and used for analysis.
For subjects experiencing shoulder pain, the intensity of pain was measured using the NPRS. An experienced physical therapist provided a detailed explanation regarding the NPRS prior to the administration of the NPRS questionnaire to ensure better understanding. All participants had to complete the NPRS questionnaire at baseline and post 4-weeks.
A sample size of 42 participants was estimated using a moderate effect size for the group x time interaction (partial η2 =0.06) (28), level of significance (α= 0.05), power of 0.80, and a 35% dropout.
Data was summarized using mean and standard deviation for quantitative variables and counts (%) for qualitative variables. The normality of continuous variables was examined using Shapiro Wilk’s test and Q-Q normality plots. The distribution of the participants’ characteristics by study group was evaluated using chi-square for qualitative variables, independent t-test for quantitative variables and Mann-Whitney test for ordinal variables. A 2-group X 3-time points (baseline, week 1, and week 4) mixed factorial analysis of variance (ANOVA) was used to examine changes in mean GH IR ROM and pain intensity (2x2, baseline vs. week 4) by study group over time.
The primary analysis included a comparison between groups using the group x time interaction effect. If the interaction was statistically significant, change from baseline was compared between groups at each follow-up time point using an independent t-test. If the results were not statistically significant, the between-groups comparison was considered not statistically significant at any time point. However, Bonferroni post hoc test was conducted on the combined groups only if the main effect of time was significant in the mixed factorial ANOVA.
The secondary analysis included testing of change from baseline at each time point within-groups using one-way repeated measures ANOVA. If the results of the test were statistically significant, post hoc comparisons using Bonferroni test was conducted on each study group separately. Partial Eta Squared (η2) was reported for significant differences in ANOVA and categorized as [0.01 small, 0.06 medium, and ≥ 0.13 large], while Cohen’s d effect size was used for independent t-test and categorized as [0.2 small, 0.5 medium, and ≥ 0.8 large] (29). The level of significance was set at p≤0.05. Statistical analysis was performed using IBM SPSS Software version 25 for Windows (Chicago, IL, USA).