Shoulder Pain and Rotational Range of Motion of the Trunk, Shoulder, and Hip in Baseball Players

The glenohumeral internal-rotation (IR) deficit and adaptive range-of-motion (ROM) gain in external rotation (ER) of the throwing shoulder is normal in baseball throwers and is characterized by an IR loss of <18° to 20° on the basis of normative data.¹ Loss of glenohumeral IR ROM may be caused by osseous adaptation and posterior shoulder tightness (16° greater humeral retrotorsion on the throwing side than on the nonthrowing side).^2–4 The total rotational ROM deficit (2.5-fold higher risk) is more predictive of a throwing-related shoulder injury than is IR ROM loss (1.9-fold higher risk).⁵ However, recent researchers⁶ determined that preseason screening of glenohumeral rotational ROM or humeral retrotorsion may not effectively identify the risk of shoulder injury in high school baseball players. Thus, the association between shoulder pain and glenohumeral IR deficit remains unclear.

Repetitive throwing volume may lead to an increase in glenohumeral rotational ROM and in hip IR ROM in the lead (nondominant) leg.⁷ During throwing, the lead leg moves the body forward, which requires more hip IR ROM; thus, the lead leg has greater hip IR ROM than the stance leg.⁷ However, studies of the effects of increased hip IR ROM on shoulder pain are lacking. Throwing a baseball involves trunk rotation in the transverse plane.⁸ Insufficient trunk rotation may result in excessive lateral leaning of the trunk during a throw, increasing stress on the glenohumeral joint.⁹ Adaptive changes in trunk rotational ROM may occur because of ROM losses or gains in the shoulder and hip.¹⁰ Healthy pitchers may show increased trunk ROM toward the nonthrowing side, which could represent compensation for a dominant-side glenohumeral IR deficit.¹⁰

No authors have examined the correlations between shoulder pain and rotational ROM of the trunk and glenohumeral and hip joints in baseball players. Previous researchers^11–15 have assessed healthy baseball players without shoulder pain and those who had a history of shoulder injury but were healthy at the time of testing. Therefore, the aims of our study were to (1) investigate the association between shoulder-pain intensity and rotational ROM of the trunk and glenohumeral and hip joints and (2) compare the differences in rotational ROM among these 3 body areas between baseball players with and those without shoulder pain.

METHODS

Procedures

Before measuring ROM, examiner A (a physical therapist), who was blinded to the other examiners, administered a questionnaire addressing shoulder-pain intensity during throwing using VAS scores, anthropometric characteristics, and the dominant arm and leg. The dominant arm was defined as the arm the player used predominantly for throwing a ball, and the ipsilateral leg was defined as the dominant leg (stance leg).

Examiners B, C, and D (ROM examiners, all physical therapists) were blinded to each other's results, to the group assignment (shoulder pain or no shoulder pain), and to each player's dominant arm. They measured the rotational ROM of the trunk, glenohumeral joint, and hip joint. All examiners had at least 3 years' experience with physical assessments and were educated on ROM measurement techniques for 1 month to improve intrarater test-retest reliability. All active rotational ROMs were measured using a standard goniometer (±1° precision) and reliable methods.^19–21 During pilot testing, we confirmed good to high intrarater test-retest reliability for all active rotational ROM values (intraclass correlation coefficient [ICC] = 0.85–0.98; standard error of measurement = 1.0°–5.0°; minimal detectable change [MDC] 90% confidence interval [CI] = 2.3°, 11.6°). Five active rotational ROM values were measured bilaterally: trunk rotation, glenohumeral IR, glenohumeral ER, hip IR, and hip ER. The order of measurements was randomly selected for the dominant versus nondominant sides and each rotational ROM evaluation using Excel (version 2016; Microsoft Corp, Redmond, WA).

Trunk rotational ROM was measured in the lunge position, which helps restrict the compensatory motion of the lumbar spine and hip during trunk rotation. It also requires stability and balance of the lower extremities during trunk rotation to remain upright, similar to that needed when throwing a ball. To measure the ROM of trunk rotation toward the dominant side, the player positioned the dominant leg forward and the nondominant leg behind the trunk with both legs in a straight line on the floor to minimize pelvic rotation.²² To measure the ROM of trunk rotation toward the nondominant side, the player moved the nondominant leg forward and the dominant leg behind the trunk. Next, the participant placed his front foot on the ground with the contralateral knee, tibia, and toe in contact with the floor. Previous authors^20,22 have demonstrated that the half-kneeling trunk-rotation test with the bar in both the front and back has good reliability. The bar-in-front position was more reliable than the bar-in-back position.²⁰ However, it is possible to protract the scapula during active trunk rotation, in which case the measurement reflects combined trunk rotation and scapular motion. To prevent compensatory scapular protraction, the participant crossed both arms over a bar in front of the chest and placed both hands on the anterior shoulder region over the bar to prevent scapular protraction during trunk rotation. The participants were asked not to compensate (ie, bend laterally and lean the trunk forward). The examiner monitored contact with the bar, the bar's parallel position to the ground, head and neck rotation, and knee and foot shifting during rotation to prevent compensation. When compensation occurred, the trial was considered a failure and repeated.²⁰ The goniometer was aligned parallel to the floor, and its axis was at the midpoint between the T1 and T2 spinous processes. The Tl and T2 spinous processes were identified by inferior palpation from the C7 spinous process, the most prominent vertebra. The location of the C7 spinous process was confirmed using the flexion-extension test, which identifies the freely moving spinous process as C6 and the stationary spinous process as C7 during active-assisted cervical flexion and extension.²³ The goniometer's stationary arm was placed parallel between the T1 and T2 spinous processes and the scapular spine opposite the rotational side. During trunk rotation, the stationary arm was held in the starting position, and the moveable arm of the goniometer followed the spine of the scapula of the opposite side from the rotational side. The player rotated his trunk toward the forward leg as much as possible while maintaining an upright posture in the lunge position and without losing his balance. He was also asked to continue looking forward at an eye-level line on the wall without holding the head and neck in the starting position while performing maximal trunk rotation.²⁴ The trunk-rotation values were then noted (Figure 1A). The average of 3 trials for each rotational ROM was used in the data analysis. The same measurement procedures were used for the dominant and nondominant sides.

View Figure

• Download as Powerpoint
• Download Figure

The examiner stabilized the scapula while assessing humeral rotational ROM on the glenoid because an isolated assessment of glenohumeral ROM (instead of the total shoulder complex) can better reflect a shoulder injury.^25,26 The procedures for active glenohumeral joint IR and ER ROM measurements were as follows. The player was asked to flex his hips and knees in a hook-lying position on the treatment table. The arm was elevated to 90° abduction in the frontal plane, and the humerus was supported with a towel to maintain neutral horizontal abduction-adduction in the transverse plane. To measure both IR and ER, the examiner placed the axis of the goniometer on the center of the olecranon and the stationary arm perpendicular to the floor and aligned the moving arm between the styloid process of the ulna and the center of the olecranon. The player then moved his humerus into either IR or ER until anterior or superior motion of the scapula was initiated, and the ROM result were recorded (Figure 1B).¹⁵ The same measurement procedures were used for the dominant and nondominant sides. The average of 3 trials for each rotational ROM was used in the data analysis. Total rotational ROM (IR + ER) of the glenohumeral joint was calculated from the sum of IR and ER ROM.

To measure the IR and ER ROM of the hip, the player sat on the edge of a table with his knees flexed to 90°. The axis of the goniometer was placed on the top of the patella, its stationary arm was aligned vertically to the floor, and its moveable arm was aligned with the tibial shaft. While the examiner stabilized the pelvis, the player performed hip IR or ER until the earliest lateral tilting of the pelvis was detected, and the value was recorded (Figure 1C).^27,28 The average of 3 trials for each rotational ROM was used in the data analysis. The same measurement procedures were used for the dominant and nondominant sides. Total rotational ROM of the hip joint was calculated from the sum of IR and ER ROM.

RESULTS

The rotational ROM data are presented in Table 1. Trunk rotational ROM values were greater in the pain group than in the control groups (mean differences = 7.0° ± 2.7° [95% CI = 1.7, 12.4] on the dominant side, P = .01, and 6.1° ± 2.7° [95% CI = 0.8, 11.5] on the nondominant side, P = .03). However, trunk ROM between the dominant and nondominant sides (P = .09) did not differ and there was no interaction of side × group (P = .82). Glenohumeral IR ROM values were lower on the dominant side than the nondominant side (mean differences = −7.9° ± 2.2° [95% CI = −12.3, −3.7] in the pain group, P = .01, and −6.4° ± 2.7° [95% CI = −11.7, −1.2] in the no-pain group, P = .02). Glenohumeral IR between the pain and no-pain groups did not differ (P = .35), and the interaction of side × group was not significant (P = .76). No differences were present in glenohumeral ER between the dominant and nondominant sides (P = .13) or groups (P = .12). Total rotational ROM between the dominant and nondominant sides (P = .06) and groups (P = .88) was not different. Hip IR ROM in the pain group was less on the dominant side than on the nondominant side, whereas no difference was demonstrated in the no-pain group (mean differences = −2.9° ± 1.3° [95% CI = −5.5, −0.3] in the pain group, P = .03, and −1.5° ± 1.7° [95% CI = −4.8, 1.9] in the no-pain group, P = .40). Hip IR between groups (P = .55) did not differ, and the interaction of side × group was not significant (P = .50). No difference was present in hip ER between the dominant and nondominant sides (P = .05) or groups (P = .97). Finally, total hip ROM did not differ between sides (P = .87) or groups (P = .72).

Table 1 Rotational Range of Motion (ROM) Measurements of Glenohumeral Joint, Trunk, and Hip Joint, Mean ± SD^a

View Fullscreen

The correlations between shoulder-pain intensity and rotational ROM for condition 1 (all players) ranged from ρ = −0.09 to ρ = 0.27 (P values = .01–.76). A significant low correlation was present between shoulder-pain intensity and trunk rotational ROM toward the dominant side (ρ = 0.27, P = .01; Table 2). The correlations between shoulder-pain intensity and rotational ROM for condition 2 (pain group only) ranged from ρ = −0.24 to ρ = 0.36 (P values = .001–0.97). A significant low correlation occurred between shoulder-pain intensity and trunk rotational ROM toward the dominant side (ρ = 0.36, P = .001; Table 2).

Table 2 Spearman Correlations Between Shoulder-Pain Intensity and Rotational Range of Motion (ROM) in All Players and in Shoulder-Pain Group Only^a

View Fullscreen

The correlations between each glenohumeral ROM (IR, ER, or total rotational ROM) and trunk or hip ROM in the shoulder-pain group ranged from ρ or r = −0.24 to 0.25 (P values = .10–.97), indicating no significant relationship (Table 3).

Table 3 Correlations Between Glenohumeral Rotational Range of Motion (ROM) and Trunk or Hip Rotational ROM in the Shoulder-Pain Group^a

View Fullscreen

DISCUSSION

The purposes of our study were to investigate the relationship between shoulder pain and rotational ROM and compare the differences in rotational ROM values of the trunk and glenohumeral and hip joints between baseball players with and those without shoulder pain. However, we found no relationship between shoulder pain and rotational ROM except for a low correlation with trunk rotational ROM toward the dominant side. Additionally, no differences were evident between the rotational ROM values of players with and those without shoulder pain, except in trunk ROM toward the dominant and nondominant sides, and the differences were clinically small. Clinicians commonly use the kinetic chain theory of energy exchange from the proximal to distal joints during assessment of ball throwing to evaluate or treat baseball players.²⁹ We noted no changes in the rotational ROM of the hip or trunk in response to glenohumeral ROM loss in adolescent players with shoulder pain; we believe ours is the first study to demonstrate this.

Although the rotational ROM values of the glenohumeral and hip joints did not differ between groups, the trunk ROM values toward the dominant and nondominant side were greater in the pain group than in the no-pain group. In addition, a significant low correlation was present between shoulder-pain intensity and trunk rotational ROM toward the dominant side (ρ = 0.27, P = .01), indicating that more trunk rotation can induce greater shoulder pain and vice versa (Figure 2). Previous authors³⁰ showed that the rotational angular velocity of the humerus was approximately 5 times that of the trunk during throwing. The greater angular acceleration of the trunk in conjunction with greater trunk ROM may result in shoulder stress.³⁰ From an injury-development perspective, a higher magnitude of trunk rotation would increase tensile loading on the soft tissues around the shoulder, such as the rotator cuff, biceps-labral complex, and joint capsule.³¹ These observations suggest that trunk rotational hypermobility may put baseball players at an increased risk of shoulder pain. However, clinicians should be cautious when interpreting the clinical relationship between shoulder-pain intensity and trunk rotational ROM, despite the differences between the shoulder-pain and no-pain groups. In our study, bilateral trunk rotational ROM values were approximately 7° higher in the pain group than in the no-pain group. This difference was higher than the MDC of half-kneeling trunk rotation at 3.82° in healthy participants based on previous data.²⁰ Yet another researcher³² reported an MDC of 10.2° in the half-kneeling position among softball players with a history of shoulder or elbow injury. Despite its statistical significance, our 7° difference failed to surpass the threshold for clinical meaningfulness (10.2°). In addition, a significant but low correlation occurred between shoulder-pain intensity and trunk rotational ROM toward the dominant side.

View Figure

• Download as Powerpoint
• Download Figure

We detected no side-to-side differences in trunk rotational ROM in either group. Contrary to our findings, earlier investigators¹⁰ described such differences in healthy baseball players: specifically, pitchers had increased trunk rotational ROM toward the nondominant side. Test position is one potential explanation for the inconsistent results between studies. The previous authors¹⁰ tested trunk rotational ROM with participants in a standing position with the trunk flexed to 90°, whereas we tested it in the lunge position. The lunge position is more similar to the position used when throwing a ball than standing with the trunk flexed and requires stability and balance in the lower extremities during trunk rotation.²⁶

Our results are consistent with those of earlier researchers^{4,15,19,33–35} who demonstrated side-to-side differences in glenohumeral IR ROM values only in healthy baseball players. Both groups showed side-to-side differences in glenohumeral IR ROM but no differences were demonstrated between groups. In addition, the side-to-side differences in our study were clinically minor (7.9° in the pain group and 6.4° in the no-pain group) compared with previous normative data on baseball throwers, which suggested that the side-to-side differences were 18° to 20°.¹ Baseball players are known to have a loss of glenohumeral IR ROM on the dominant side of 18° to 20° due to humeral retrotorsion and posterior shoulder tightness, which is not pathologic.¹ For hip IR ROM, the side-to-side difference was 2.9° in the pain group, which is not clinically important.¹⁵ Hence, we suggest that side-to-side differences in these body areas may not be associated with shoulder pain but rather may be characteristic of all baseball players, regardless of shoulder pain.

Previous authors¹⁵ suggested the need for further studies to evaluate the relationship between glenohumeral IR and trunk and hip ROM in players with shoulder pain. However, based on our results, we found no relationship between glenohumeral and trunk or hip ROM in the shoulder-pain group (Table 3). Thus, we suggest that no clinical association exists between glenohumeral IR deficit and hip ROM change, regardless of shoulder pain in baseball players.

CLINICAL IMPLICATIONS

We observed no clinical correlations between shoulder-pain intensity and the rotational ROM values of the trunk and glenohumeral and hip joints. These findings do not mean that baseball players do not have a pathologic restriction of the trunk or hips, because we did not differentiate whether ROM restrictions were less than normal limits. Therefore, pathologic restriction of the trunk and hip may still cause shoulder ROM restriction and pain. When developing management plans for baseball players with shoulder pain, sports clinicians should consider multidirectional factors that may contribute to shoulder pain in baseball players, such as pathologic restriction, competition level, pitch volume, and throwing frequency rather than evaluating only joint ROM.

LIMITATIONS

The first limitation of this study is that our players were adolescents; as a result, it is difficult to generalize our results to professional baseball players. Second, we did not subdivide the groups according to specific shoulder injuries, such as rotator cuff tears or superior-labrum anterior and posterior lesions. A third limitation was the unbalanced numbers of participants in each group. More players were in the pain group (60 players) than in the no-pain group (35 players). With regard to the statistical analysis, we could not fully compare ROM differences between groups because of the unbalanced number of participants. Thus, a future comparison study with equal distributions of participants in the pain and no-pain groups is needed. And finally, when measuring thoracic rotational ROM, it is difficult to maintain the goniometer's stationary arm while correctly placing the moveable arm. Earlier investigators²⁰ described good reliability using this method to measure thoracic rotational ROM (within-session interrater reliability ICC = 0.94, within-day intrarater reliability ICC = 0.92). However, no researchers have demonstrated the validity of radiography for this purpose. Further studies are needed to determine the validity of thoracic rotational ROM measured using a goniometer versus radiography.

CONCLUSIONS

The purposes of our study were to investigate the association between shoulder-pain intensity and rotational ROM of the trunk and glenohumeral and hip joints and compare the differences in rotational ROM values of these 3 body areas between baseball players with and those without shoulder pain. However, no clinical correlations were noted between shoulder-pain intensity and the ROM of each body area. Additionally, no clinically meaningful differences were present in the ROM values of these body areas between players with and those without shoulder pain. Thus, we did not find rotational ROM changes of the hip and trunk in response to a loss of glenohumeral ROM in adolescent players with shoulder pain.