Throwing Shoulder Adaptations Are Not Related to Shoulder Injury or Pain: A Preliminary Report

Abstract

Background:

Overhead throwing in baseball and softball athletes induces shoulder adaptations theorized to increase risk of shoulder musculoskeletal injury (MSI) and/or pain due to range of motion (ROM) deficits.

Hypothesis:

Shoulder ROM adaptations are associated with a higher risk for developing shoulder MSI and pain.

Study Design:

Prospective cohort study.

Level of Evidence:

Level 3.

Methods:

A total of 60 National Collegiate Athletic Association Division I athletes cleared for full athletic participation and free from upper extremity MSI in the last 4 weeks (age, 19.0 ± 1.2 years; weight, 82.1 ± 13.7 kg; height, 178.6 ± 11.2 cm; softball, n = 23; baseball, n = 37). Passive glenohumeral internal rotation (IR), external rotation (ER), and horizontal adduction (HA) ROM were measured with the scapula stabilized and used to categorize participants with/without external rotation gain (ERG), external rotation insufficiency (ERI), glenohumeral internal rotation deficit (GIRD), pathological GIRD, and posterior shoulder tightness (PST) before the competitive season. Groups were then compared to assess the incidence of shoulder MSI prospectively and prevalence of shoulder pain at the initial evaluation.

Results:

Baseball and softball athletes demonstrated significantly less IR ROM in the dominant shoulder (50.6° ± 9.4°) compared with the nondominant shoulder (59.1° ± 8.6°; P < 0.01) and significantly more ER ROM (dominant, 104.6° ± 12.1°; nondominant, 97.7° ± 12.0°; P < 0.01). Incidence of shoulder MSI was 15% but was not significantly related to any shoulder adaptations. No significant relationship was found between prevalence of pain and any shoulder adaptations in the 27% of athletes with pain.

Conclusion:

Increased ER and decreased IR ROM adaptations in intercollegiate overhead throwing athletes do not appear to be correlated to risk of shoulder MSI or pain.

Clinical Relevance:

The findings of this level 3 prospective study provide clinicians working with overhead athletes information regarding shoulder MSI risk and pain. It is recommended that clinicians should not use ROM adaptations exclusively to determine increased risk of shoulder MSI.

Shoulder musculoskeletal injury (MSI) persists in overhead athletes, such as baseball and softball players, despite years of research on preventative measures. Boltz et al ¹ reported 16.01% of injuries occurred at the shoulder in collegiate baseball athletes, whereas Veillard et al ³³ reported 15.2% of shoulder injuries in collegiate softball athletes. In addition, the throwing mechanism accounted for 61% of shoulder injuries in the same population of softball players. ³³ The shoulder complex in overhead throwing athletes endures increased stress due to the mechanical demands of throwing. ³ Musculoskeletal adaptations have been theorized to be induced by these forces during overhead throwing^5,11,12; however, it is still unknown if these adaptations increase or mitigate the risk of shoulder MSI. ¹³

Shoulder range of motion (ROM) adaptations are thought to result from forces generated during the cocking and deceleration phases of overhead throwing. During the cocking phase, to achieve maximum velocity, the athlete must achieve a greater range of shoulder external rotation (ER) ROM to maximize the arc of rotation. ³ This increases the shoulder internal rotation (IR) forces required to propel the ball forward. ³ During deceleration, the posterior musculature (rotator cuff) contracts eccentrically, potentially causing microtrauma to the muscles,^2,10 which result in scar tissue formation and stiffness.^10,11 Increased ER ROM and eccentric overload of the posterior musculature in the shoulder complex during throwing are theorized to be causative factors for shoulder ROM adaptations.^2,3,11,21

Adaptive ROM changes typically observed in overhead athletes include external rotation gain (ERG), external rotation insufficiency (ERI), glenohumeral internal rotation deficit (GIRD), pathological GIRD, and posterior shoulder tightness (PST). ERG is defined as an increase of ≥5° of shoulder ER ROM in the dominant compared with the nondominant arm, and is expected due to the repetitive hyper-ER of throwing.^10,11 Conversely, ERI is a decrease of ≥5° in shoulder ER ROM and, when present in the dominant arm, is thought to increase MSI risk.^10,11 An increase in ER and a decrease in IR ROM in the throwing shoulder of overhead athletes has been identified in previous studies that results in a shift of the total range of motion (TRM) in the throwing arm.^2,7,11,38 This shift in TRM as compared bilaterally should not be considered pathologic but rather a sport-specific adapatation.^11,38 Potential MSI risk may come when there is a deficit in TRM. ¹¹ GIRD, defined as a decrease of ≥18° of glenohumeral IR in the dominant arm compared with the nondominant arm,^15,36 contributes to the TRM deficit when it exceeds the gain in ER ROM. Burkhart et al ³ defined this as pathological GIRD.^25,28,36,37 During the pitching or throwing motion, the repetitive eccentric overload amid deceleration is the overuse mechanism associated with PST, which may be assessed with horizontal adduction (HA) ROM. HA ROM of <10° has been shown to be significantly related to increased forces in the upper extremity during throwing, ¹⁸ however, a threshold for HA that increases MSI risk has not yet been determined. ¹¹

Literature is conflicted as to whether these ROM adaptations are correlated with prospective MSI.^{3,11,13,14,21,31,36,37} Further, it is unknown if there is a relationship between shoulder ROM adaptations, specifically with intercollegiate baseball and softball athletes, and the presence of shoulder pain. Given the culture of sports and mentality of competitive athletes, some participants who are fully cleared for participation may be playing with shoulder pain. This is supported by the findings by Frisch et al ⁸ that the incidence of shoulder pain was twice as high as the incidence of shoulder injury in high school volleyball athletes. The presence of pain, therefore, was not an exclusion criterion in in the current study as the aim was to collect data on overhead athletes with no medical restrictions to sports participation.

The purposes of this study were to (1) evaluate the incidence of shoulder MSI and its relationship to specific shoulder adaptations in the dominant arm and (2) evaluate the prevalence of pain in throwing athletes cleared for full sports participation and its relationship to specific shoulder adaptations in the dominant arm. It was hypothesized that all shoulder adaptations would be associated with prospective shoulder MSI as well as the prevalence of pain. While research is still inconclusive as to the relationship between these adaptations and the development of shoulder MSI and prevalence of pain, it is clear these adaptations alter the normal function of the shoulder, which has the potential to increase risk of injury and development of pain. The findings of this prospective study could provide clinicians working with overhead throwing athletes additional knowledge relative to shoulder MSI risk to incorporate when designing and implementing injury prevention programs.

Methods

A prospective study design was utilized to evaluate the relationship between musculoskeletal characteristics and the incidence of shoulder MSI and pain prevalence. Ethical approval for this study was granted by an Institutional Review Board, and written informed consent was obtained for each participant before data collection. To participate in this study, participants were required to be between 18 and 35 years old; currently rostered and cleared for full physical activity as an intercollegiate baseball or softball player; and free from any upper extremity MSI within the last 4 weeks. Participation criteria and detailed data collection procedures have been previously published. ³² All data, with the exception of prospective MSI, were collected in a single session before the competitive season.

Shoulder ROM

Bilateral passive shoulder ROM measurements included glenohumeral IR, ER, and HA using previously established methods, and were measured to the nearest degree using a digital inclinometer (GHIR/GHER) and goniometer (HA).^22,24 Testing order was standardized for all participants: glenohumeral IR, ER, and HA. For all measures, the participant was supine and the scapula was stabilized to isolate glenohumeral motion. Passive ROM measurements were performed by a coinvestigator (athletic trainer or physical therapist) who passively moved the limb to the point of end-feel, with the measurement obtained by a second coinvestigator, ³² neither of whom were blinded to recorded ROM value. Intratester reliability for these measures by the evaluators in the current study ranged from 0.840 to 0.910 and intertester reliability ranged from 0.723 to 0.864. ³² To evaluate the effect of shoulder ROM adaptations on prospective MSI, raw measures were used to create the following categorical variables: PST, GIRD, ERG, ERI, and pathological GIRD as defined in Table 1.

Table 1.

Definition of shoulder range of motion adaptations

Variable	Definition	Source
PST	Passive horizontal adduction of <10°	Laudner et al 17
GIRD	A decrease in the dominant shoulder’s passive internal rotation of ≥18° compared with the nondominant shoulder’s passive internal rotation	Wilk et al 33 and Kibler et al 14
ERG	An increase in the dominant shoulder’s passive external rotation of >5° compared with the nondominant shoulder’s passive external rotation	Hellem et al 10 and Harp et al 9
ERI	A decrease in the dominant shoulder’s passive external rotation of <5° compared with the nondominant shoulder’s passive external rotation	Hellem et al 10 and Harp et al 9
Pathological GIRD	The presence of both GIRD and TRM deficit in the dominant shoulder	Burkhart et al 3

ERG, external rotation gain; ERI, external rotation insufficiency; GIRD, glenohumeral internal rotation deficit; PST, posterior shoulder tightness; TRM, total range of motion.

Pain

Current pain level in the throwing (dominant) shoulder only was collected using a questionnaire (Figure 1) at the beginning of the laboratory data collection session before the season. Collecting pain data at a single timepoint before competition allowed the investigators to identify the point prevalence of pain within the sample of participants. Participants were ensured that their answers would be kept confidential and not affect playing status.

Figure 1.

Excerpt from the current health questionnaire given to participants to fill out.

Prospective Injury

Shoulder MSI data were collected throughout each participant’s active participation in the study to determine shoulder MSI incidence rates. An injury incidence rate is defined as the occurrence of a “new injury” in a given population relative to the total population at risk during a specified period of time. To be consistent with previous literature, this study utilized the National Collegiate Athletic Association (NCAA) Injury Surveillance System definition of injury. ⁴ At the end of each season, a coinvestigator met with the team athletic trainer, who provided shoulder MSI data from each participant’s medical chart, including mechanism of injury, onset, injury type, location, sublocation, and time lost. Any participant with an incident MSI was categorized into the injured group, whereas participants with no “new” MSI were categorized as uninjured. Incidence proportion (number of new MSI/total participants) was calculated.

Statistical Analysis

Data analyses were completed using SPSS for Windows software (Version 27.0; SPSS, Inc). Descriptive statistics were calculated for dominant and nondominant shoulder ROM characteristics. Continuous data were assessed for normality using Shapiro-Wilk tests. Paired t tests and effect size were utilized to assess for statistical differences between dominant and nondominant shoulder ROM characteristics. For categorical data, frequencies were used to establish point prevalence of pain and incidence proportion for MSI. Two-tailed Fisher exact tests and odds ratios (OR) were utilized to evaluate the association between shoulder ROM adaptations and prospective shoulder MSI as well as shoulder pain. The criterion for significance was established at alpha = 0.05, a priori.

Results

Over the course of the 5 competitive seasons, a total of 74 athletes volunteered and met the criteria to participate in the study; 10 participants were excluded from data analysis due to incomplete datasets and 4 softball pitchers were excluded due to differences between the windmill pitch and overhead pitch/throwing motion. A total of 60 participants across all positions (excluding softball pitchers) were included in data analysis. Participant demographics can be found in Table 2, and a breakdown of participants by sport and position is presented in Table 3.

Table 2.

Subject demographics

Variable	Overall	Softball Players	Baseball Players
Number, N	60	23	37
Age, y	19.0 ± 1.2	19.2 ± 1.3	19.6 ± 1.0
Weight, kg	82.1 ± 13.7	68.4 ± 9.2	90.6 ± 7.8
Height, cm	178.6 ± 11.2	166.6 ± 5.1	186.0 ± 6.4

Table 3.

Participant positions

	N
Baseball (total)	23
Catcher	6
Infield	5
Outfield	6
Pitcher	20
Softball (total)	37
Catcher	8
Infield	6
Outfield	9
Pitcher	0

Shoulder ROM: Bilateral Comparisons

Descriptive statistics for ROM characteristics and statistical outcomes for bilateral asymmetries are provided in Table 4. Dominant shoulders had significantly less IR ROM (P < 0.01) and significantly greater ER ROM (P < 0.01) compared with nondominant shoulders. There was no statistically significant difference in TRM, with average TRM <1.5° on both the dominant and nondominant sides. When analyzed by sport, both baseball and softball players demonstrated significantly less glenohumeral IR ROM (both P < 0.01) and significantly greater glenohumeral ER ROM (P < 0.03, P < 0.01, respectively) in the dominant shoulder as compared with the nondominant shoulder. A significant difference in HA ROM was identified only in softball players, who demonstrated significantly less HA ROM (P = 0.02) on the dominant shoulder as compared with the nondominant side. Despite statistical significance, the mean difference for HA ROM was only 2.3°, which does not reach the minimal detectable change of 9° previously recommended to exclude variability or measurement error.^16,17 Further, the average HA in the dominant shoulder within both sports did not meet the threshold of <10° HA to qualify as PST.

Table 4.

Comparison between dominant and nondominant shoulder ROM in throwing athletes

		Dominant Shoulder	Nondominant Shoulder		Effect Size
Parameter	N	Mean (°) ± SD	Mean ± SD	P Value	Point Estimate	Lower Bound	Upper Bound
Overall
HA	60	17.8 ± 8.7	18.9 ± 8.7	0.21	-0.162	-0.416	0.093
Glenohumeral internal rotation a	60	50.6 ± 9.4	59.1 ± 8.6	<0.01	-0.892	-1.189	-0.59
Glenohumeral external rotation a	60	104.6 ± 12.1	97.7 ± 12.0	<0.01	0.666	0.384	0.944
TRM	60	155.3 ± 14.5	156.8 ± 13.5	0.28	-0.141	-0.395	0.114
Baseball
HA	23	18.3 ± 9.5	17.9 ± 8.2	0.60	0.112	-0.3	0.52
GIRD a	23	54.3 ± 8.6	60.4 ± 7.5	<0.01	-0.747	-1.204	-0.277
Glenohumeral external rotation a	23	104.3 ± 11.5	99 ± 10.5	0.03	0.486	0.048	0.914
Total ROM	23	158.5 ± 8.8	159.4 ± 8.8	0.74	-0.07	-0.478	0.34
Softball
HAD a	37	17.5 ± 8.2	19.8 ± 9	0.02	-0.415	-0.748	-0.076
GIRD a	37	48.4 ± 9.3	58.3 ± 9.2	<0.01	-0.989	-1.379	-0.59
Glenohumeral external rotation a	37	104.9 ± 12.5	96.9 ± 12.9	<0.01	-0.978	-1.364	-0.583
TRM	37	153.2 ± 15.7	155.2 ± 15.7	0.25	-0.193	-0.517	0.134

GIRD, glenohumeral internal rotation deficit; HA, horizontal adduction; ROM, range of motion; TRM, total ROM.

^aSignificant difference between dominant and nondominant shoulder ROM at the level of α = 0.05.

Prevalence of Shoulder Pain

At the time of data collection, 27% of participants (N = 16 of 59) reported some continuous or intermittent shoulder pain, with the 19% (11 of 59) reporting pain after activity. A pain score was not received from 1 participant; therefore, 59 participants were included for analysis that included pain. Specific percentages for self-reported shoulder pain categories are presented in Table 5. When examining point prevalence of pain by sport, 22% (N = 8 of 36) of softball players and 35% (N = 8 of 23) of baseball players reported some amount of shoulder pain, with the highest percentage reporting pain after activity (19% and 17%, respectively). Pain prevalence by sport is presented in Table 6.

Table 5.

Pain in the dominant shoulder

	N (%)
Self-reported pain
No pain	43 (73)
Pain before activity	2 (3)
Pain after activity	11 (19)
Pain before and after activity	1 (2)
Pain before, during, and after activity	2 (3)

Table 6.

Pain point prevalence by sport

	N	%
Baseball	23
No pain	15	65
Yes pain	8	35
Before activity	1	4
After activity	4	17
Before and after activity	1	4
Before, during, and after activity	2	9
Softball	36
No pain	28	78
Yes pain	8	22
Before activity	1	3
After activity	7	19
Before and after activity	0	0
Before, during, and after activity	0	0

Incidence Proportion of Shoulder MSI

Throughout the duration of the study, 9 MSI occurred in the dominant shoulder: an incidence proportion of 15%. Type of injuries are presented in Table 7.

Table 7.

Injury demographics

	Prospective Injuries
Type
Bursitis	1
Bicep tendonitis	2
Impingement	3
Nerve injury	1
Sprain	2
Cause
Throwing	7
Landing	1
Other	1

Shoulder Adaptations and Pain

There were no significant associations between the presence of any shoulder adaptation and shoulder pain (Table 8). The point prevalence of shoulder pain was 3% (2 of 59) for participants classified as having PST, compared with 24% (14 of 59) for those without PST (P = 0.66). The prevalence of shoulder pain was 5% (3 of 59) for participants classified as having GIRD, compared with 13% (13 of 59) for those without GIRD (P ≥ 0.99). The prevalence of shoulder pain was 2% (2 of 59) for participants classified as having ERI, compared with 24% (14 of 59) without ERI (P ≥ 0.99). The prevalence of shoulder pain was 15% (9 of 59) for participants with ERG, compared with 12% (7 of 59) for those without ERG (P = 0.77). The prevalence of shoulder pain was 3% (2 of 59) for participants with pathological GIRD, compared with 24% (14/59) without pathological GIRD (P = 0.71). Calculated ORs showed participants classified as having any of the shoulder adaptations had similar odds of shoulder pain as those classified without shoulder adaptations (Table 8).

Table 8.

Association of frequency of shoulder pain in athletes with musculoskeletal adaptations

	Shoulder Pain		Fisher Exact Test		95% CI
Adaptation	Negative (N = 43)	Positive (N = 16)	P Value	OR	Lower Bound	Upper Bound
Posterior shoulder tightness
Negative (N = 53)	39 (66%)	14 (24%)	0.66	1.393	0.229	8.459
Positive (N = 6)	4 (7%)	2 (3%)
Isolated GIRD
Negative (N = 47)	34 (58%)	13 (22%)	≥0.99	0.872	0.204	3.734
Positive (N = 12)	9 (15%)	3 (5%)
Isolated ER insufficiency
Negative (N = 50)	36 (61%)	14 (24%)	≥0.99	0.735	0.136	3.975
Positive (N = 9)	7 (12%)	2 (3%)
Isolated ER gain
Negative (N = 23)	16 (27%)	7 (12%)	0.77	0.762	0.238	2.443
Positive (N = 36)	27 (46%)	9 (15%)
Pathological GIRD
Negative (N = 49)	35 (59%)	14 (24%)	0.71	0.625	0.118	3.316
Positive (N = 10)	8 (14%)	2 (3%)

ER, external rotation; GIRD, glenohumeral internal rotation deficit; OR, odds ratio.

Shoulder Adaptations and Incidence of Shoulder MSI

No significant associations were found between the presence of any shoulder adaptation and incidence of shoulder MSI in this study (Table 9). The incidence proportion of shoulder MSI was 0% (0 of 60) for participants classified as having PST compared with 15% (9 of 60) for those without PST (P = 0.58). The incidence proportion of shoulder MSI was 2% (1 of 60) for participants classified as having GIRD compared with 13% (8 of 60) for those without GIRD (P = 0.67). The incidence proportion of shoulder MSI was 3% (2 of 60) for participants classified as having ERI compared with 12% (7 of 60) for those without ERI (P = 0.61). The incidence proportion of shoulder MSI was 8% (5 of 60) for participants classified as having ERG compared with 7% (4 of 60) for those without ERG (P = 0.72). The incidence proportion for shoulder MSI was 2% (1 of 60) for participants classified as having pathological GIRD compared with 13% (8 of 60) for those without pathological GIRD (P ≥ 0.99). Calculated ORs showed participants classified as having any of the shoulder adaptations had similar odds of sustaining a shoulder MSI as those classified without shoulder adaptations (Table 9).

Table 9.

Association of frequency of prospective shoulder injury in athletes with musculoskeletal adaptations

	Prospective Shoulder Injury		Fisher’s Exact Test		95% CI
	Negative (N = 51)	Positive (N = 9)	P Value	OR	Lower Bound	Upper Bound
Posterior shoulder tightness a
Negative (N = 54)	45 (75%)	9 (15%)	0.58	–	–	–
Positive (N = 6)	6 (10%)	0 (0%)
Isolated GIRD
Negative (N = 48)	40 (67%)	8 (13%)	0.67	0.455	0.051	4.034
Positive (N = 12)	11 (18%)	1 (2%)
Isolated ER insufficiency
Negative (N = 51)	44 (73%)	7 (12%)	0.61	1.79	0.308	10.462
Positive (N = 9)	7 (12%)	2 (3%)
Isolated ER gain
Negative (N = 23)	19 (32%)	4 (7%)	0.72	0.742	0.177	3.108
Positive (N = 37)	32 (53%)	5 (8%)
Pathological GIRD
Negative (N = 50)	42 (70%)	8 (13%)	≥0.99	0.583	0.065	5.265
Positive (N = 10)	9 (15%)	1 (2%)

ER, external rotation; GIRD, glenohumeral internal rotation deficit; OR, odds ratio.

^aOR cannot be calculated due to a prevalence of 0 in the positive PST group.

Discussion

The identification of specific shoulder adaptions within overhead athletes, specifically throwing athletes, has sparked concern as to whether these adaptations are protective or a risk factor for shoulder MSI. The purpose of this study was to evaluate the incidence of shoulder MSI and the frequency of pain and injury in throwing athletes with and without specific shoulder adaptations in the dominant arm. It was hypothesized that shoulder adaptations would be associated with a presence of shoulder pain and prospective shoulder MSI. However, based on the results of this study, when not grouped by sport or position, the hypothesis was rejected as no significant differences in self-reported pain or incidence of shoulder MSI were found when comparing participants with the presence of specific shoulder adaptations to those without. The current study provides preliminary insight as to the longitudinal relationship shoulder adaptations may have on the risk of developing shoulder MSI as well as pain point prevalence. Using the findings of this study, clinicians can be better informed as to the potential risk of shoulder MSI for their throwing patient when these shoulder adaptations are identified.

Shoulder ROM

The findings of this study were consistent with some previous research that found overhead throwing athletes possess adaptations to the throwing shoulder, including decreased glenohumeral IR and increased ER ROM while maintaining a similar total rotational ROM as compared bilaterally.^5,7,19 More specifically, decreased glenohumeral IR and increased ER ROM have been reported in the throwing shoulder of baseball players^5,7; however, no statistical differences in IR ROM, ER ROM, or TRM have been reported in softball players.^{26 -28} Previous research with softball players has utilized both position players and pitchers, which may explain the discrepancies between previous studies with baseball or softball participants as well as with the current study. In the current study, softball pitchers were excluded intentionally due to the difference in mechanics of the windmill pitch, as compared with the overhead throwing motion/baseball pitch, and the effect these difference mechanical demands could have on soft tissue adaptations. In a study evaluating shoulder adaptations in both baseball and softball players, Hibberd et al ¹² reported that baseball position players demonstrated significantly greater GIRD compared with softball players and controls, but similar ERG compared with softball players. Based on the general agreement with previous literature, it may be reasonable to infer that there are similarities in overhead throwing athletes (ie, baseball and softball position players) in terms of soft tissue adaptations around the glenohumeral joint that may not be present in softball pitchers due to differences in biomechanical demand.

Between and within participants for both sports, there was a significant decrease in IR and increase in ER ROM in the dominant shoulder compared with the nondominant side; however, there was no difference in TRM. This could indicate the presence of humeral retroversion, which is a nonmodifiable musculoskeletal characteristic. Humeral retroversion is a bony adaptation of humeral head angle often seen in throwing athletes, but is not synonymous with GIRD or TRM, ³⁴ which can contribute to a loss in isolated shoulder IR and an increase in ER in the throwing arm while maintaining a comparable total arc of motion relative to the nonthrowing arm. ³⁵ Given the methods of the current study, the influence of humeral retroversion on the findings cannot be ascertained. The current study aimed to utilize clinically friendly measures to assess potential modifiable characteristics; therefore, no diagnostic imaging (eg, CT scans) were obtained. Future research should evaluate the role of humeral retroversion in the presence of various shoulder adaptations as well as its relationship to pain and MSI.

Shoulder Adaptations and Pain

Of the throwing athletes who participated in this study, 27% (16 of 60) reported some amount of shoulder pain related to sport participation upon initial evaluation. Although there is limited research assessing pain in softball and baseball players, no studies have specifically examined pain in baseball and softball athletes relative to ROM deficits. The findings of the current study are, however, consistent with prevalence rates for pain reported in athletes participating in other overhead sports. In elite overhead athletes combined, the point prevalence for shoulder pain was reported as 21.4%, ²⁰ while 36% elite female handball players reported current shoulder pain. ²³ Although the previous studies included overhead athletes, none included softball or baseball players, which may help explain the slight variations in prevalence.

No shoulder adaptations were found to be associated with the prevalence of shoulder pain. It is plausible that, after years of repetitive throwing, the throwing athlete becomes accustomed to the ROM and mechanical demands of throwing and that these adaptations are not inherently painful. Numerous research studies have reported increased ER and decreased IR ROM adaptations in the dominant arm of healthy, pain-free throwing athletes.^{2,3,5-7,12,29,36} This may explain why no relationships between shoulder pain and shoulder adaptations were found in the current study. Other factors have been identified that may contribute more directly to pain in overhead athletes. Mohseni-Bandpei et al ²⁰ found that pain in swimming, handball, volleyball, basketball, rowing, and wrestling athletes was correlated to sex, body mass index, years of practice, level of sport, satisfaction of income, and days of practice per week. Collectively, these findings indicate that other internal and external factors beyond sport-specific shoulder adaptations may contribute to the development of shoulder pain in overhead athletes.

Further, pain is subjective and dependent on factors such as awareness, fear, and concentration ³⁰ ; therefore, athletes may underreport self-perceived pain as it has been suggested that athletes are more likely to ignore and suppress pain in fear of lost playing time. ³⁰ Hall et al ⁹ reported that, while athletes experienced more pain on a day-to-day basis from training and games compared with the general population, athletes have a higher pain tolerance compared with nonathletes.

In the current study, the shoulder adaptation with the highest prevalence of shoulder pain was in participants with ERG (19%). Although no statistically significant association was found, of the 15 athletes who reported shoulder pain, 11 (73%) also had ERG. Marcondes et al ¹⁹ found that painful shoulders in tennis players had significantly more ERG compared with asymptomatic shoulders. Conversely, Myklebust et al ²³ found no significant differences in ROM between painful and nonpainful shoulders. Collectively, these studies provide insight that shoulder pain in overhead athletes should not be ignored as it is likely present in these athletes, even when they are cleared for full participation in their sport. Due to the conflicting findings, additional investigation is warranted in the relationship between ERG and pain.

It does not appear that the prevalence of shoulder pain and ROM adaptations in intercollegiate baseball and softball athletes have previously been examined concurrently. The participants in the current study were considered healthy and fully cleared for participation in practice and competition; however, over one-quarter of the athletes reported having some amount of shoulder pain. While mild, self-limiting pain may be expected during periods of increased tissue loading (eg, preseason, return to sport participation), there is also a culture in sport to hide and play through pain to protect starting positions and playing time. More research is needed to assess the prevalence of pain and its association and effect on injury and health in overhead athletes. Since the current study assessed pain at one point in time and did not assess pain severity, future research should examine whether changes in pain over time are related to changes in shoulder ROM as well as the association between pain severity and injury and health.

Shoulder Adaptations and Incidence of Shoulder MSI

Categorical criteria for rotational adaptations do not appear to be a predictor of prospective shoulder MSI. The findings of the current study are consistent with more robust and higher levels of evidence studies that have identified a lack of predictive value of these shoulder adaptations. A recently published meta-analysis, ¹⁴ which identified and pooled prospective studies aiming to determine risk of shoulder MSI, found incidence of upper extremity injury was not associated with GIRD, TRM, TRM loss, glenohumeral ER, or glenohumeral ERG. ¹⁴

In the current study, GIRD was defined to be a deficit of ≥18°, based on the recommendations of Kibler et al ¹⁵ and the 2012 Throwing Summit. This study found that the presence of GIRD was not predictive of shoulder MSI in throwing athletes, which was consistent with the findings of several previous studies.^28,36 Multiple studies have identified IR ROM deficits in injured throwing athletes compared with noninjured throwing athletes^29,36; however, the incidence of injury was not statistically significant between pitchers classified as having GIRD and those who did not. ³⁶ In contrast, several studies have identified a relationship between GIRD and presence of injury in overhead athletes.^13,28 Due to methodological differences, there are limitations to the application and generalization of these findings across studies. Across the literature, there is a lack of standardized classification of GIRD. This is probably due to the fact that traditional classifications have not resulted in conclusive findings. Further, several studies looked at incidence of upper extremity injury, which included both shoulder and elbow injuries, while the current study only evaluated association with shoulder injuries. Therefore, these findings, which identified an association of decreased IR and incidence of injury, may contribute more directly to injury at the elbow, which was not investigated in our study.

The presence of ERI and ERG and their association on shoulder MSI were evaluated in the current study and no significant association was found with either ERI or ERG and the incidence of shoulder injury. These findings were consistent with Shitara et al, who also found no significant difference in ER ROM in injured high school baseball pitchers compared with noninjured baseball pitchers. ²⁹ The majority of previous literature has reported TRM rather than isolated ER, which makes it difficult to directly compare results across studies. TRM deficit has been defined as a decrease of >5° in TRM compared bilaterally.^25,36 In the current study, it was postulated that this could be redundant of the presence of GIRD, in that TRM does not identify where the deficit is occurring. It is feasible that a throwing athlete could have an ER ROM deficit or gain without an IR ROM deficit (or vice versa) that would result in a difference in TRM bilaterally. Empirically, it is suggested that, to maintain a similar TRM bilaterally, the TRM shifts in the throwing shoulder with a similar gain in ER ROM that accounts for the deficit in IR ROM, and that the deficit in TRM is due to a lack of gain in ER ROM. The combined loss of both IR and ER ROM is what some considered to be the definition of “pathological GIRD.”^3,25 Clearer terminology of adaptations and consistent use is needed regarding GIRD, pathological GIRD, and TRM deficits, and the influence of ER ROM on the defined deficit will have direct implications on affected tissue and intervention strategies. While measuring TRM may be a good starting point, it is likely necessary to evaluate IR and ER ROM in isolation as well.

When creating categorical groups for PST adaptations, the presence of PST was not found to be a predictor of prospective shoulder MSI. Categorical groups for PST adaptations were based on Laudner et al. ¹⁸ Six athletes (10%) were identified as having PST, with none sustaining a shoulder MSI. These findings were consistent with those of Shitara et al, ²⁹ who found no association between PST and prospective upper extremity MSI. Shanely et al ²⁸ evaluated HA ROM and found that baseball and softball players demonstrated less HA ROM in the dominant shoulder compared with the nondominant side, and that those athletes who sustained an upper extremity injury demonstrated significantly lower HA ROM compared with noninjured athletes.

Clinical Recommendations

The results of this study display limited quality of patient-oriented evidence with inconsistencies across the literature. However, supported clinical knowledge on shoulder MSI and pain in relation to overhead athletes is presented. The Strength of Recommendation Taxonomy is a B for this evidence, as ROM adaptation should not be the sole factor to establish shoulder MSI risk; however, ROM adaptations may be taken into consideration when implementing preventative measures and rehabilitation techniques.

Limitations

For this longitudinal study, a power analysis was not conducted; therefore, given the small sample with regard to injuries, there is a likely risk of type II error, which may account for the lack of significant findings and is a recognized major limitation of this study. It was appropriate to combine baseball position players, pitchers, and softball position players and this may limit the application of these findings to specific sports. Due to low incidence of injury, it was also necessary to include all shoulder injuries; this may limit significant findings as some shoulder adaptations may be related more directly to specific shoulder injuries. Although shoulder MSI are listed based on specific tissues involved (Table 7), identification of the origin of shoulder pain based on physical examination or imaging is difficult and not accounted for in this study. In addition, only shoulder MSIs that resulted in time lost from sport participation were included. This presents as a major limitation because an athlete who may have developed pain and/or a diagnosed injury that did not result in time away from sport was therefore not included. The definition of injury used in the current study is, however, consistent with the NCAA Injury Surveillance System and allows for standardization and comparison with other published literature. ⁴ The lack of blinding of the evaluator to the ROM measurements is a limitation; however, given that this was a prospective cohort study, knowing the ROM at the time of testing would not contribute to bias based on prospective injury. This study did not assess for humeral torsion, which could impact ROM measurements.

Conclusion

The study determined that an increase in ER ROM and a deficit in IR ROM are expected adaptations to the throwing shoulder of baseball and softball players, excluding pitchers, potentially due to the demands of the throwing mechanism. Moreover, no association between ROM adaptations and prospective shoulder injury, or between ROM adaptations and pain prevalence, was found. However, this may be due to inadequate participant numbers. In addition, clinicians should not use ROM adaptations exclusively to determine increased risk of shoulder MSI in this population.