The Efficacy of Upper-Extremity Elastic Resistance Training on Shoulder Strength and Performance: A Systematic Review

Elastic resistance exercise is a popular mode of strength training that has demonstrated positive effects on whole-body strength and performance. The purpose of this work was to identify the efficacy of elastic resistance training on improving upper limb strength and performance measures for the shoulder. Seven online databases were searched with a focus on longitudinal studies assessing shoulder elastic training strength interventions. In total, 1367 studies were initially screened for relevancy; 24 full-text articles were included for review. Exercise interventions ranged from 4–12 weeks, assessing pre-/post-strength and performance measures inclusive of isometric and isokinetic strength, 1RM strength, force-velocity tests, and throwing-velocity tests. Significant increases in various isometric strength measures (IR:11–13%, ER:11–42%, FL: 14–36%, EXT: 4–17%, ABD: 8–16%), 1RM strength (~24% in bench press), force-velocities, throwing- and serve-velocities (12%) were all observed. Elastic resistance training elicited positive effects for both strength and performance parameters regardless of intervention duration. Similar significant increases were observed in isometric strength and 1RM strength across durations. Isokinetic strength increases were variable and dependent on the joint velocity conditions. Quantifying the dosage of appropriate exercise prescription for optimal strength and performance gains is inconclusive with this study due to the heterogeneity of the intervention protocols.


Introduction
Strength training provides a multitude of health and performance benefits. This training involves a diverse range of movements that require the muscles to counteract some form of resistance or force. The use of this training method enables improvements in static and dynamic muscle function, bone strength and formation, joint range of motion, joint stability [1], and athletic performance while decreasing injury risk [2][3][4]. These same benefits hold true across the lifespan, with demonstrated improvements in both young and old participant populations [1,[3][4][5][6]. While many strength training modalities are available, differences exist in the volume of research across modalities and the efficacy of their long-term use as a training methodology.
Elastic resistance training (ERT) is a popular method of resistance training. This type of training enables users to perform functional movements in any direction, alternative to traditional free weights, which provide an external force for the muscles against gravity. Upper extremity ERT typically involves attaching the band to a wall or post, standing at a distance to create tension in the band while performing rectilinear, curvilinear, or circular motions with each arm. This can facilitate increases in maximum torque production and stabilization at the shoulder [5,7]. ERT has sport-specific utility; overhead athletes have used elastic resistance to mimic throwing motions including external rotation and abduction movements during warmup [8]. Continued research on elastic resistance has elaborated on muscle activation differences, particularly in posture examinations and the use of singleand dual-vector elastic setups [9,10].
ERT is a viable mode of resistance training to employ in strengthening and rehabilitation programs due to its low cost, versatility, and applicability to all populations. This method of resistance training has received support for its simplicity and feasibility among elderly populations [11,12], effective home-based programs [13,14], and for its usefulness in advanced lifters and athletic populations by providing a varied form of resistance [15,16]. Elastic bands have diverse training applications, including speed and agility training, stretching, plyometric training, and reactive neuromuscular training [17]. The simplicity, versatility, and inexpensiveness of elastic bands for training could combat the commonly perceived barriers to strength training such as the fear of injury, high costs of training equipment, and the intimidation of using fitness facilities [18].
The effects of this type of training on whole-body muscle strength have previously been explored [19,20], but the efficacy of this specific type of training on isokinetic and isometric strength measures of the shoulder remains unknown. Strength gains observed with single-joint and multi-joint elastic resistance exercises have shown to be comparable to that of conventional resistance training [19,[21][22][23][24]. The effects of resistance training have been documented to be affected by the sex, health status, and initial strength capability of the user, and should be considered in the context of this treatment method. A comprehensive analysis of elastic training and its effects on shoulder strength would provide clinicians, rehabilitation specialists, and strength and conditioning coaches with information to determine its utility for their clients. The purpose of this review was to assess the current literature on elastic resistance training and collate its findings to determine its efficacy on shoulder strength and performance parameters.

Materials and Methods
A search was devised by considering the main topics of interest and carefully selecting keywords to efficiently extract articles from each database. Seven databases in total were searched on 11 December 2020, including MEDLINE, Embase, Web of Science, PubMed, ProQuest Dissertations and Theses, SPORTDiscus and CINAHL. The search strategy was critiqued and revised by the institutional Library staff to formulate a finalized search string ( Figure 1). The search was comprised of a combination of three classifications with their affiliated keywords; these sections were focused on the upper extremity, strength, and performance measures, and the elastic resistance training modality. The study was registered to PROSPERO (ID: CRD42021236849).
All articles were screened for eligibility to be included within this review. Included were randomized control trials, systematic reviews/meta-analyses, cohort studies, and theses. No restrictions on publication dates were applied. A minimum training regimen of twice weekly for 4 weeks with pre-and post-regimen strength or performance metrics was required for randomized control trial inclusion. The strength and performance parameters could include isokinetic or isometric strength assessments, one-repetition maximum (1RM) testing, force-velocity tests, or throwing-velocity tests. The participant population was limited to healthy subjects of all ages. Exclusion criteria involved incorrect study design, patient populations with shoulder pathologies or known adverse health conditions, or outcome measures that did not assess pre-and post-muscular strength.
A multi-step screening process was applied to arrive at the final selection of studies for full-text analysis and data extraction. Initial database searches were conducted by the primary author and were subsequently extracted and screened according to the inclusion/exclusion criteria. The articles were then uploaded to Covidence (Veritas Health Innovation Ltd., Melbourne, Australia) to manage screening processes. Abstract and title screening was conducted first by two independent reviewers, with a third independent reviewer to resolve reviewer decision conflicts. Following the initial screening, eligible articles were assessed using a full-text screen with identical criteria and conflict resolution methods as the first round ( Figure 2). All articles that passed the second round of screening were the final studies included for data extraction. Search string entries for all databases. * represents a wildcard operator to increase the possible number of search terms that contain the preceding letters (e.g. 'shoulder' and 'shoulders' are both returned by shoulder *). N2 means that the words appear within two words of each other.

Figure 2.
All titles collected through database searches were screened for eligibility; article dispersion is described through this process flowchart.
A modified Downs and Black quality checklist [25] was used to assess the methodological quality of each study that was extracted; fourteen criteria from this checklist were utilized. Each study was given a score for each criterion with "yes" = 1, "no/unable to determine" = 0, and a total score out of 14 was yielded. The threshold for adequate quality was a score of 7; any article with a score lower than this was excluded from the review. The ratings of each article are provided in Table 1. All studies were assessed for potential risks of bias by using the Risk Of Bias In Nonrandomised Studies [47] ( Table 2). The level of bias assigned to each study was formulated from seven domains: The randomization process, intervention deviations, absence of outcome data, measurements of outcomes, and the reported selection of results. Escalating ratings were described as low (L), moderate (M), serious (S), and critical (C). Bias ratings were totaled based on the highest rating within the seven categories.    The principal summary measure extracted from all included articles was the differences in pre-and post-test mean strength and performance measures. These measures were divided into five main categories, with strength-based measures including isokinetic strength, isometric strength, and 1RM strength, as well as two performance measures of throwing-velocity and force-velocity tests. A percent-change was calculated from the pre-/post-strength and performance measures to normalize increases or decreases of the parameters being measured that occurred over the strength training protocol. This allowed for quantification in strength and performance gains and uniformity of the different variables of performance and strength measurements.

Study Selection
Collectively, 1367 studies were extracted through initial database retrieval; 24 articles were extracted for assessment following the screening. A detailed flowchart of article screening is detailed in Figure 2. Each article included in the review had data extraction through six key components deemed essential for appropriate analysis of this strength intervention, which included the participant pool, study duration, elastic training intervention exercises, session details, mode of strength measurement, and strength quantification pre-and post-intervention. A full table and the characteristics for each respective study can be found in Appendix A.

Bias and Quality Assessments
The methodological quality of each study was represented by a score out of 14, and an overall bias rating was given for each article. The methodological quality scores and the rating of bias for each article are located in Tables 1 and 2, respectively. Qualitative scores for all studies were no less than 7, with a range of 7 to 12 (50-86%) and a mean score of 10 [13]. Risk of bias assessments of each article conducted using the ROBINS-I identified one study with serious risk of bias due to deviation from the intended interventions [27]; five studies had a moderate risk of bias in multiple domains, including concerns due to confounding [44], participant selection [7], intervention classification [44][45][46], deviations from intended interventions [41], and missing data [7,45].

Exercise Protocol
The duration of the elastic-training interventions ranged from 4 to 12 weeks. Five studies were 4 weeks in duration [7,30,31,36,38,44], nine studies were 6 weeks in duration [27,34,39,40,42,43,48], four studies were 8 weeks in duration [26,41,44,45], and four studies were comprised of 10-12-week interventions [28,33,37,46,49]. The majority of intervention protocols were three days per week (78%), as one study consisted of two exercise sessions each week [26], and one study with a protocol of five days per week [30]. The most commonly employed elastic resistance bands in the studies assessed in this review were Theraband ® at varying resistance levels. Many studies made use of multiple colourresistance levels as a source of progressive overload over the course of the intervention. Overload was also introduced by increasing band stretch to facilitate increased tension. Three studies did not specify the type of banded resistance used [37][38][39] and one study used an "MVP band", a circular band that attaches around the wrist rather than being held in the hand [31].
The exercises completed and the angles of resistance of the band were variable. The bands used in nearly all exercise protocols were fixed to a wall or doorknob at hip height or elbow height. Many studies did not specify band tension at the onset of exercise; those that did, started either in a slack or high-tension setting. Specific exercises were generally described as classifications of movements, including abduction exercises, shoulder-retraction exercises, flexion and extension exercises, and internal and external rotation exercises. Appendix A provides specific exercise-session details including repetitions, sets, and the number of sessions per week.

Strength Performance Assessment Results
Various strength and performance measures were assessed within the included studies. Handheld dynamometers and isokinetic dynamometers were the most popular method of measuring upper-extremity strength in the studies included (70%), which measure muscle strength through isometric contraction or specific joint-velocity conditions [50,51]. The most common isokinetic dynamometers used were the Kin-Com ® or CYBEX ® and were typically employed at joint velocities from 60-240 • per second [7,28,39,48,49]. Performance and strength measures collected using 1RM tests included variations of lying bench press, dumbbell pullover, seated row, shoulder press, and shoulder abduction [26,33,43,45]. Forcevelocity tests were performed on a Monark cycle ergometer [26] and throwing-and servingvelocity evaluations were conducted with the use of a radar gun [26,31].
Time-dependent measures are illustrated in Appendix A. Increases in various strength and performance measures were observed regardless of the length of the exercise intervention. Significant increases of 7-42% improvement were observed in multiple isometric strength measures across all studies. Studies measuring internal and external rotation isometric strength observed significant increases in both parameters, with increases in internal and external rotation isometric strength increasing by 11.2-13.5 and 11.0-42.3% across studies, respectively [29,30,38,44]. Increases in isometric flexion and extension were observed across all studies, with significant (p < 0.05) increases of 14.7-36.0% and 4.7-17.1%, respectively. [30,34,35,37,52]. The few studies that assessed isometric abduction observed both significant (p < 0.05) [30,37] and insignificant increases [41,52] ranging from 8-16%. All studies that evaluated 1RM strength found increases that were significant, regardless of the duration of the strength training program [26,43,45]. An average increase of 24% was observed for lying and seated variations of the bench press. Richards and Dawson [43] found significant increases of 11.4-25.2% in 1RM shoulder flexion and abduction strength after six weeks of training, as other studies observed increases in various concen-tric lifts [26,33,45]. Throwing and serving velocity increases were observed over variable intervention durations. Escamilla et al. [31] found significant increases in baseball-throwing velocities (3.9%) after a four-week protocol, as significant increases in both force-velocities and throwing-velocities (W peak increases of 36%) were observed by Aloui et al. [26].
The changes that occurred for isokinetic strength were much more variable over time. Significant increases were observed by Baker [27] and Batalha [28] in external rotation at 180 degrees/s (4.2-4.4%), while Page et al. [48] observed decreases of 14.8% during isokinetic diagonal movement patterns at 180 • /s. Decreases in strength were observed at lower internal rotation joint velocities (60 • /s) by Baker and Batalha, with decreases of 2.1-2.6%. Similar decreases (2.7-4.6%) were found in eccentric internal and external rotation strengths after eight weeks by Sugimoto et al. [44]. Opposingly, Treiber [7] and Knerr [36] observed increases in all joint velocity conditions for both external and internal rotation (2.3-21.2%).

Discussion
The primary aim of this review was to examine the effects of longitudinal ERT programs on shoulder strength and performance measures. Although the effects of elastic resistance training on whole-body strength gains have previously been explored, the specific effects of this mode of training on the upper extremity remained unclear. A consensus from the included studies identified statistically significant increases in external rotation, internal rotation, flexion and extension, and abduction strength of the shoulder following varied ERT programs. Secondly, this mode of strength training yielded significant increases in 1RM strength-particularly in the lying and seated bench press-and performance measures including throw and serve velocities that increased by~11% over a six-to eight-week regimen [26,31,32]. Lastly, there are some positive effects of ERT on isokinetic strength, though, these results are less conclusive due to varying observations at different joint velocities. Some studies observed increases at all joint velocities [27,36] and increases in both eccentric and concentric ER and IR strengths [36,40,44] while some studies observed decreases in certain variables of isokinetic strength, such as Page [48] and Mascarin [40] who observed decreases at higher joint velocities at 180 • /s and 240 • /s, respectively. Collectively, the increases observed in performance measures and strength variables, particularly in peak torque, 1RM strength, force velocities, and throwing velocities identify ERT as a viable mode of resistance training for eliciting observable strength and performance gains in individuals participating in longitudinal strength training programs. These 1RM increases were observed across a diverse participant pool; the five studies that had 1RM as an outcome metric included university-aged participants of both sexes [33], national-level handball players [26], and post-menopausal women [45]; these groups collectively saw reported 1RM increases of 13-24%. These 1RM tests were completed using non-elastic equipment such as CYBEX machines, but strong relationships between submaximal elastic resistance and estimated maximal strength have been quantified and could be used in future research designs [53]. Due to the lack of homogeneity in the exercise interventions of the studies included and the range of initial strength and normalized strength increases, an optimal prescription of upper extremity training with elastic resistance cannot be concluded, and further research is needed.
Previous studies have concluded that elastic devices utilized in training can produce strength gains that are equivalent to those observed with free weights and conventionaldevice training [19,54]. ERT has also previously been proven to be effective in improving whole-body strength and function in the elderly [3,55]. This is similar to the findings within this review, as a minimum of a four-week elastic resistance training program has demonstrated improvements in isometric strength and 1RM strength, in addition to other performance measures and isokinetic strength variables. The ease of access and versatility of this type of training paired with the positive effects observed among studies included in this review indicated this training may be useful for clinicians and trainers to implement as a longitudinal program to aid in the strength and performance of their clients.
There are limitations to be considered for the intervention methods employed. The strength training protocols for studies included in this review were not uniform. The intervention protocols varied considerably in length, where strength and performance gains could be attributed to different factors. The increases observed for shorter intervention durations were likely due to neurological adaptations such as increased motor unit recruitment and neural drive to the working muscles, whereas longer intervention durations likely elicited strength gains from both central and peripheral adaptations [56,57]. As muscle volume was not measured in the examined studies, the noted strength gains could have resulted from a combination of factors. There is a paucity of documented strength changes with elastic resistance training in populations beyond healthy, university-aged participants, despite the simplicity and feasibility of its use [1,3,11]. The papers included in this review focused on strength. Thus, studies with a rehabilitation focus that may not have documented a strength outcome were not included. This work highlights the need to continue to examine the effects of this training method on expanded populations, including those with differences in initial strength or health, as well as understanding differences across sex and age groups. Variability existed in the exercises completed and their accompanying sets and repetitions, making inferences regarding appropriate dosing of an ERT regimen to elicit optimal strength and performance gains difficult. The starting tension of the elastic band was inconsistent across studies, as some programs instructed participants to begin with minimal resistance or tension, while others had participants initiate movements at a distance from the wall that removed slack in the band. Discrepancies between the slack length or length of stretch could be a cause for patient variability and a lack of codified information on resistance levels within these studies obscure recommendations. The initial resting length of the bands is a crucial component to consider, as the resistance that is generated throughout any movement completed with the band is dependent upon the relative stretch of the material. Many of the intervention protocols implemented in each study did not provide details on the monitoring of exercises being completed, the set-up of the bands, or the progression of the coloured resistance levels over time. While progression over the training period occurred in most included studies, it was unclear what percentage of the maximum the elastic training represented. Lastly, variability existed in the methodology of measuring strength variables, particularly in isometric and isokinetic strength [58]. These strength measures were assessed at different internal and external rotation angles, and different joint velocities ranging from 60-240 • /s, which may have confounded results.

Conclusions
Longitudinal elastic resistance training protocols involving upper extremity movements such as external rotation, internal rotation, abduction, and elevation elicit increases in strength and performance for the general healthy population. The considerable heterogeneity of the exercise regimens and methods of assessing strength make it difficult to firmly conclude the types of exercises and protocols that should be employed in training and clinical settings to elicit the most observable strength and performance enhancements. The documented changes in strength may represent a portion of the progressive changes seen in these groups, and additional reviews focusing on the potential existence of technique, range of motion, or fatigue resistance changes when using this exercise methodology would be beneficial. This enforces the need for more rigorous studies that follow a more standardized exercise regimen and protocol of measuring strength and performance parameters.

Acknowledgments:
The authors would like to thank Chelsea Humphries for her contributions to search strategies for this work.

Conflicts of Interest:
The authors declare no conflict of interest.
Appendix A Table A1. Summary of studies extracted, including performance measures, the sample population, type and length of training, exercise-session details, measurements of strength, and the quantification in pre-/post-strength and performance.