Influence of glenohumeral internal rotation deficit on shoulder rotation torque in adolescent baseball players: An isokinetic comparative cross-sectional study

Abstract

Background

Glenohumeral internal rotation range of motion deficit (GIRD) is linked to shoulder strength in the glenohumeral (GH) joint and to increased risk of throwing-related injuries. In adolescent overhead athletes, GIRD ≥25° has been associated with a four-fold higher risk of upper-extremity injury. However, from a functional perspective relevant to overhead throwing, the extent to which GIRD is related to the dynamic strength of the shoulder rotators has not been investigated.

Objective

This study compared internal rotation (IR) and external rotation (ER) torque between adolescent baseball athletes with and without GIRD, and examined factors associated with concentric and eccentric IR/ER torque, including GH joint range of motion (ROM).

Method

Eighty male adolescent baseball athletes participated. IR, ER, total rotational ROM, and GIRD were measured. Concentric and eccentric torque of IR and ER, as well as functional strength ratios, were assessed. Group differences were analyzed with independent t-tests, correlations with Pearson's coefficients, and predictors with Stepwise multiple regression analysis.

Results

Concentric IR and ER torques were significantly lower in the GIRD group than in the non-GIRD group (mean differences: 4.6% and 6.9% body weight, respectively; p < 0.05). Concentric torques showed poor to fair correlations with GH joint rotational ROM, particularly reduced IR ROM and increased GIRD. Eccentric torques did not differ significantly between groups. Regression analysis identified IR ROM and GIRD as significant predictors of concentric torque, while antagonist exhibited the greatest relative influence on the agonist.

Conclusions

Concentric shoulder strength (IR and ER torques) is reduced in adolescent baseball athletes with GIRD and is significantly associated with IR ROM loss. IR ROM and GIRD were key predictors of concentric torque, while antagonist strength exerted the greatest relative influence on agonist function. In contrast, eccentric torques showed no group differences or meaningful correlation with GH joint ROM.

Keywords

shoulder rotation range of motion isokinetic torque glenohumeral internal rotation deficit adolescent baseball players

1 Introduction

Shoulder pain is the most prevalent concern among overhead athletes, with more than 50% of adolescent baseball players experiencing shoulder pain.^1,2 Recent epidemiological studies have reported a notable prevalence of shoulder and elbow injuries in youth and high school baseball players.^3,4 The glenohumeral (GH) joint is particularly susceptible to microtrauma due to repetitive throwing. Internal rotation (IR) velocities of approximately 7000–7900°/s and torques of up to 90 Nm have been reported.^1,5,6 Consequently, distractive forces and eccentric stress increase on the soft tissues of the posterior shoulder, including the posterior deltoid, rotator cuff, and posteroinferior capsule.^7–9 In addition, the excessive external rotation (ER) motion required during the late cocking phase leads to posterior–superior humeral head translation, thereby further increasing mechanical stress on the GH joint.^7,10,11 Recent imaging and clinical reviews have confirmed that repetitive posteroinferior capsular contracture, scapular dysfunction, and microinstability exacerbate posterosuperior internal impingement and predispose throwers to SLAP lesions and rotator cuff pathology.^6,11

These biomechanical features of overhead throwing may induce adaptive changes in the soft tissues. Glenohumeral internal rotation deficit (GIRD), defined as a reduction in IR range of motion (ROM) of the throwing shoulder, is considered a representative adaptive response to repetitive throwing-induced soft tissue stress.^7,12 Typically, an IR ROM reduction of more than 20° in the throwing shoulder compared to the non-throwing shoulders is the suggested criterion for GIRD.^13–15 Some studies, however, have proposed that a GIRD of up to 15° may be considered within the normal range in overhead athletes.^16,17 Importantly, GIRD exceeding 20° is closely associated with throwing-related injuries,^15,18–20 with reports indicating approximately a four-fold increase in upper-extremity injury risk in adolescent overhead athletes presenting with GIRD of ≥25°.¹⁹ Complementing these findings, a systematic review and meta-analysis across multiple overhead sports further suggested that athletes with GIRD tend to show an increased risk of shoulder and elbow injuries, particularly in association with total rotational motion deficits and excessive ER gain,²¹ which may represent a compensatory mechanism to maintain functional performance despite a reduced total rotational arc.^14,22

Furthermore, GIRD is associated with alterations in shoulder muscle strength, as decreased ROM affects muscle-length tension relationships and subsequently influences force generation.²³ In addition, GIRD shifts the GH rotation center posterosuperiorly, increasing distractive forces on the posterior GH structures.^7,13 These biomechanical alterations contribute to imbalances in the shoulder rotator muscles,⁷ and previous studies have demonstrated significant differences in IR and ER strength, rotator thickness, and strength ratios in adolescent and adult overhead athletes with GIRD.^20,24–26 During throwing, the GH joint experiences maximal stress owing to repetitive concentric contractions of the ER and IR agonists, particularly from the late cocking to deceleration phases.^9,27,28 Therefore, eccentric antagonist control is essential to prevent GH joint overload.^29,30

The functional strength ratio (FSR), defined as the ratio of eccentric antagonist strength to concentric agonist strength, has been proposed as an indicator of shoulder stability.^31–34 Supporting this concept, Guney et al.²⁶ found that adolescent overhead athletes with GIRD showed reduced FSR, while recent scoping reviews of isokinetic dynamometry in youth athletes, including baseball players, have consistently reported ER/IR strength imbalances, particularly reduced eccentric ER (ECC ER) to concentric IR (CON IR) ratios.³⁵ A lower ECC ER/CON IR ratio has been suggested as a predisposing factor for throwing-related injuries in overhead athletes.^22,35,36 Although some assumptions posit that GH joint rotational ROM and GIRD influence shoulder muscle strength,^7,13,23 few studies have explore how GH rotational ROM affects the concentric and eccentric torques of the IR and ER muscles, which reflect the functional characteristics required for throwing.

Previous investigations have primarily relied on isometric strength assessments, which may not adequately capture the dynamic demands of overhead throwing.^20,24,25 Evidence from adolescent athletes indicates that isokinetic dynamometry can be reliably applied in youth athletes, with reliability confirmed in adolescent athletes³⁷ (ICCs 0.72–0.99, CVs 3.9–17.4%) and further studies profiling isokinetic strength characteristics in adolescent and junior overhead athletes,^26,33 together with a recent scoping review that examined its broader application in youth athletes.³⁵ However, comprehensive evaluation of shoulder rotational strength in adolescent overhead athletes remains scarce, with, to our knowledge, only one investigation directly examining isokinetic strength in relation to GIRD.²⁶ By contrast, scoping reviews have systematically mapped and synthesized existing evidence on ER/IR strength characteristics in youth overhead athletes, yet these reviews did not specifically evaluate their relevance to GIRD.³⁵ Therefore, this study aimed to investigate differences in the concentric and eccentric torques of the IR and ER muscles and the FSR according to GIRD status in adolescent baseball players, and to examine potential factors associated with CON and ECC IR and ER torque, including GH joint ROM.

2 Materials and methods

2.1 Subjects

Sample size was estimated using G*Power (v3.1.2; University of Kiel, Germany). A pilot study with 30 male adolescent baseball athletes was conducted, and correlation analyses were performed between IR ROM and concentric/eccentric IR and ER torques. The lowest correlation coefficient observed was between IR ROM and ECC IR torque (r = −0.186), which was used as the effect size for the sample size estimation. Based on this value, a priori power analysis (correlation: bivariate normal model, α = 0.05, power = 0.95) indicated that a minimum of 53 participants was required. Accordingly, 80 male adolescent baseball players who were able to train and compete without pain were recruited. Inclusion criteria required a Kerlan-Jobe Orthopedic Clinic (KJOC) score ≥70, which is above the pain threshold of 68.6.³⁸ Demographic data are presented in Table 1. Participants with a history of shoulder surgery, dysfunction, or upper- extremity injuries were excluded. Informed consent was obtained from all parents or guardians, and consent for participation was obtained for all study participants. The study was approved by the Bioethics Committee of Pukyong National University (Approval No: 1041386-202310-HR-118-02) and registered with the Clinical Research Information Service (KCT0009349).

Table 1.
Demographic and clinical characteristics of participants (n = 80).

Variables Descriptive statistics, mean (SD) Range

Age (years) 17.21 (1.06) 15 to 19

Height (cm) 176.72 (5.92) 161.2 to 195

Weight (kg) 73.68 (8.20) 59.1 to 94.9

Baseball experience (years) 6.00 (1.31) 2 to 9

KJOC score 84.47 (3.95) 73.9 to 92

Training characteristics

U/E weight frequency/week 3.28 (0.71) 3.28 to 5

U/E weight time (hour)/week 5.43 (1.08) 5.43 to 8

U/E stretching frequency/week 6.29 (0.46) 6 to 7

U/E stretching time (min)/week 89.13 (20.73) 60 to 140

Pitching or throwing frequency/week 5.80 (0.70) 5 to 7

Pitching or throwing time (hour)/week 15.08 (2.52) 10 to 21

Batting frequency/week 4.78 (1.94) 1 to 7

Batting time (hour)/week 10.39 (5.33) 1.50 to 21

IR ROM (°), mean (SD) 67.74 (10.93) 45 to 88

ER ROM (°), mean (SD) 98.71 (8.38) 82 to 115

Total ROM (°), mean (SD) 166.44 (17.06) 142 to 183

GIRD (°), mean (SD) 11.13 (7.73) 0 to 30

GIRD grade, N (%) GIRD of <15°, 39(48.8%)

GIRD of 15° - 20°, 36(45.0%)

GIRD of ≥20°, 5(6.3%)

CON IR torque (%BW) 37.07 ± 6.78 19.73 to 52.87

CON ER torque (%BW) 39.18 ± 7.97 24.40 to 53.97

ECC IR torque (%BW) 40.67 ± 5.10 29.33 to 53.80

ECC ER torque (%BW) 42.21 ± 6.34 27.10 to 55.20

Variables	Descriptive statistics, mean (SD)	Range
Age (years)	17.21 (1.06)	15 to 19
Height (cm)	176.72 (5.92)	161.2 to 195
Weight (kg)	73.68 (8.20)	59.1 to 94.9
Baseball experience (years)	6.00 (1.31)	2 to 9
KJOC score	84.47 (3.95)	73.9 to 92
Training characteristics
U/E weight frequency/week	3.28 (0.71)	3.28 to 5
U/E weight time (hour)/week	5.43 (1.08)	5.43 to 8
U/E stretching frequency/week	6.29 (0.46)	6 to 7
U/E stretching time (min)/week	89.13 (20.73)	60 to 140
Pitching or throwing frequency/week	5.80 (0.70)	5 to 7
Pitching or throwing time (hour)/week	15.08 (2.52)	10 to 21
Batting frequency/week	4.78 (1.94)	1 to 7
Batting time (hour)/week	10.39 (5.33)	1.50 to 21
IR ROM (°), mean (SD)	67.74 (10.93)	45 to 88
ER ROM (°), mean (SD)	98.71 (8.38)	82 to 115
Total ROM (°), mean (SD)	166.44 (17.06)	142 to 183
GIRD (°), mean (SD)	11.13 (7.73)	0 to 30
GIRD grade, N (%)	GIRD of <15°, 39(48.8%)
GIRD of 15° - 20°, 36(45.0%)
GIRD of ≥20°, 5(6.3%)
CON IR torque (%BW)	37.07 ± 6.78	19.73 to 52.87
CON ER torque (%BW)	39.18 ± 7.97	24.40 to 53.97
ECC IR torque (%BW)	40.67 ± 5.10	29.33 to 53.80
ECC ER torque (%BW)	42.21 ± 6.34	27.10 to 55.20

Abbreviation: KJOC, Kerlan-Jobe Orthopedic Clinic Shoulder and Elbow score; U/E, upper extremity; IR, internal rotation; ER, external rotation; ROM, range of motion; GIRD, glenohumeral internal rotation deficit; CON IR, concentric internal rotation; CON ER, concentric external rotation; ECC IR, eccentric internal rotation; ECC ER, eccentric external rotation; BW, body weight.

2.2 Measurement of GH joint ROM

GH joint ROM was assessed through IR, ER, total rotational ROM, and GIRD. IR and ER were passively measured at 90° shoulder abduction using a bubble goniometer by two examiners, one stabilizing the scapula and the other recording ROM.^15,39,40 The goniometer axis was aligned with the olecranon process; the stationary arm was positioned perpendicular to the floor, and the movable arm was aligned parallel to the ulna's styloid process (Figure 1). GIRD was calculated as the difference in IR between throwing and non-throwing shoulders, while total rotation was defined as the sum of IR and ER of the throwing shoulder. Intra-rater reliability was assessed using ICC (3,1) according to COSMIN guidelines. The same examiner performed repeated ROM measurements twice, separated by a 5-day interval, under identical testing conditions. The results demonstrated high reliability for IR (ICC = 0.925, 95% CI = 0.886–0.951, SEM = 3.0, p < .001) and ER (ICC = 0.926, 95% CI = 0.887–0.952, SEM = 2.3, p < .001). Participants with GIRD ≥15° were assigned to the GIRD group (n = 41), and those with <15° were assigned to the non-GIRD group (n = 39).

Figure 1.

Measurement of internal and external rotation ROM in the glenohumeral joint. (a) Internal rotation, (b) external rotation.

2.3 Measurement of isokinetic torques of the IR, ER, and FSR

The isokinetic concentric and eccentric torques of IR and ER were assessed using the Biodex System 4 (Biodex Corp., Shirley, NY). Participants completed 3–5 submaximal warm-up contractions under standardized conditions. All measurements were conducted in the supine position, which provides stability, facilitates ROM assessment, and minimizes compensatory movements.^41,42 Participants were positioned with 90° shoulder abduction, 0° rotation, and 90° elbow flexion in forearm pronation. The shoulder axis was aligned with the dynamometer, and straps were used to secure the trunk and shoulder girdle (Figure 2). Torque was measured across IR 60° to ER 90°^31,32 at a velocity of 120°/s.^30,43 Each participant performed five repetitions, and peak torque values were normalized to body weight (%BW). FSRs were calculated using ECC IR/CON ER and ECC ER/CON IR ratios to reflect the dynamic balance between agonist and antagonist rotator muscles. Two types of ratios were considered: (1) ECC IR/CON ER, representing the functional demand during the cocking phase; and (2) ECC ER/CON IR, representing the functional demand during the deceleration phase. These ratios have been described in previous studies as important indicators of dynamic stability in athletes, including overhead athletes.^34,44 A 5-min rest interval was provided between tests. Gravity correction was omitted because torque was generated in the direction of gravity regardless of contraction type.^32,45

Figure 2.

Measurement of internal and external rotation ROM in the glenohumeral joint. (a) Internal rotation, (b) external rotation.

2.4 Statistical analyses

Data were analyzed using SPSS version 21.0. The normality of all variables was assessed using Kolmogorov-Smirnov test (p > 0.05). For group comparisons, independent t-tests were performed separately for each dependent variable (concentric and eccentric IR/ER torque and FSR). Each dependent variable was tested only once between the two groups (GIRD vs. non-GIRD). Pearson's correlation coefficients were used to examine relationships between torque values, GH joint ROM, and other variables (e.g., physical characteristics, experience, and upper limb training and practice routines). Finally, stepwise multiple regression analysis were conducted to identify significant predictors of concentric and eccentric IR/ER torque.

3 Results

3.1 Difference in concentric and eccentric torques of the IR and ER, and FSR between GIRD and non-GIRD

Table 2 presents torque comparisons between GIRD and non-GIRD groups. CON IR torque was significantly lower in the GIRD group (mean difference: 4.61 PT/BW, effect size [ES]: 0.72), as was CON ER torque (mean difference: 6.88 PT/BW, ES: 0.95). The ECC IR/CON ER ratio was significantly higher in the GIRD group (mean difference: −0.19, ES: 0.70). No significant differences were observed in eccentric torques or in the ECC ER/CON IR ratio.

Table 2.
Difference in concentric and eccentric torques of IR and ER, and FSR between GIRD and non-GIRD groups.

Group

Variables GIRD (n = 41) Non-GIRD (n = 39) Mean difference 95% CI p

CON IR torque (%BW) 34.83 ± 6.37 39.44 ± 6.45 −4.61 −1.76, −7.46 0.002*

CON ER torque (%BW) 35.83 ± 7.66 42.71 ± 6.74 −6.88 −3.67, −10.10 <0.001*

ECC IR torque (%BW) 40.25 ± 5.66 41.11 ± 4.47 −0.86 −3.14, 1.41 0.452

ECC ER torque (%BW) 41.57 ± 5.87 42.88 ± 6.82 −1.31 −4.13, 1.52 0.360

ECC IR/CON ER ratio 1.17 ± 0.31 0.99 ± 0.19 0.19 0.07, 0.30 0.002*

ECC ER/CON IR ratio 1.23 ± 0.29 1.13 ± 0.31 0.10 −0.28, 0.24 0.121

	Group
CON IR torque (%BW)	34.83 ± 6.37	39.44 ± 6.45	−4.61	−1.76, −7.46	0.002*
CON ER torque (%BW)	35.83 ± 7.66	42.71 ± 6.74	−6.88	−3.67, −10.10	<0.001*
ECC IR torque (%BW)	40.25 ± 5.66	41.11 ± 4.47	−0.86	−3.14, 1.41	0.452
ECC ER torque (%BW)	41.57 ± 5.87	42.88 ± 6.82	−1.31	−4.13, 1.52	0.360
ECC IR/CON ER ratio	1.17 ± 0.31	0.99 ± 0.19	0.19	0.07, 0.30	0.002*
ECC ER/CON IR ratio	1.23 ± 0.29	1.13 ± 0.31	0.10	−0.28, 0.24	0.121

Abbreviation: IR, internal rotation; ER, external rotation; FSR, functional strength ratio; GIRD, glenohumeral internal rotation deficit; 95% CI, 95% confidence interval; CON IR, concentric internal rotation; CON ER, concentric external rotation; ECC IR, eccentric internal rotation; ECC ER, eccentric external rotation; BW, body weight ECC IR/CON ER, eccentric internal rotation to concentric external rotation; ECC ER/CON IR, eccentric external rotation to concentric internal rotation.

*p < 0.05.

3.2 The results of analysis of the correlation

Table 3 summarizes the correlation results. CON IR torque demonstrated fair positive correlations with IR ROM (r = 0.417) and total rotation ROM (r = 0.239), and fair negative correlations with ER ROM (r = −0.306) and GIRD (r = −0.444). It also showed a fair positive correlation with CON ER torque (r = 0.619). CON ER torque showed a fair positive correlation with IR ROM (r = 0.390), and fair negative correlations with ER ROM (r = −0.418) and GIRD (r = −0.510). Additionally, CON ER torque demonstrated a poor positive correlation with baseball experience (r = 0.231), as well as fair positive correlations with CON IR (r = 0.619) and ECC ER torque (r = 0.433). ECC IR and ECC ER torques showed no significant correlations with GH joint ROM variables. ECC IR torque demonstrated poor negative correlations with height (r = −0.248) and weight (r = −0.291), and a fair positive correlation with ECC ER torque (r = 0.454). ECC ER torque also showed fair positive correlations with CON ER torque (r = 0.433) and ECC IR torque (r = 0.454).

Table 3.
Pearson correlation coefficients (r) among variables (n = 80).

IR and ER torques

CON IR torque (%BW) CON ER torque (%BW) ECC IR torque (%BW) ECC ER torque (%BW)

Correlation coefficient Correlation coefficient Correlation coefficient Correlation coefficient

IR ROM 0.417 0.390 −0.003 0.057

ER ROM −0.306* −0.418** −0.097 −0.188

Total ROM 0.239* 0.091 −0.101 −0.114

GIRD −0.444 −0.510 −0.040 −0.084

CON IR torque (%BW) 0.619 0.078 −0.012

CON ER torque (%BW) 0.619 0.131 0.433

ECC IR torque (%BW) 0.078 0.131 0.454

ECC ER torque (%BW) −0.012 0.433 0.454

Height (cm) −0.093 −0.093 −0.248* 0.006

Weight (kg) −0.125 0.004 −0.291* −0.198

Baseball experience (years) 0.116 0.231* −0.110 −0.096

KJOC score −0.130 −0.068 −0.015 −0.050

U/E weight training frequency −0.012 −0.024 0.077 0.115

U/E weight training time (hour) 0.190 0.168 0.063 −0.016

U/E stretching frequency −0.122 −0.077 −0.007 −0.070

U/E stretching time (min) 0.101 0.047 −0.072 0.055

Pitching or throwing training frequency 0.077 0.135 −0.152 0.067

Pitching or throwing training time (hour) 0.031 0.131 −0.094 0.164

Batting training frequency 0.048 −0.113 0.187 −0.063

Batting training time (hour) 0.057 −0.098 0.186 −0.030

	IR and ER torques
IR ROM	0.417**	0.390**	−0.003	0.057
ER ROM	−0.306*	−0.418**	−0.097	−0.188
Total ROM	0.239*	0.091	−0.101	−0.114
GIRD	−0.444**	−0.510**	−0.040	−0.084
CON IR torque (%BW)		0.619**	0.078	−0.012
CON ER torque (%BW)	0.619**		0.131	0.433**
ECC IR torque (%BW)	0.078	0.131		0.454**
ECC ER torque (%BW)	−0.012	0.433**	0.454**
Height (cm)	−0.093	−0.093	−0.248*	0.006
Weight (kg)	−0.125	0.004	−0.291*	−0.198
Baseball experience (years)	0.116	0.231*	−0.110	−0.096
KJOC score	−0.130	−0.068	−0.015	−0.050
U/E weight training frequency	−0.012	−0.024	0.077	0.115
U/E weight training time (hour)	0.190	0.168	0.063	−0.016
U/E stretching frequency	−0.122	−0.077	−0.007	−0.070
U/E stretching time (min)	0.101	0.047	−0.072	0.055
Pitching or throwing training frequency	0.077	0.135	−0.152	0.067
Pitching or throwing training time (hour)	0.031	0.131	−0.094	0.164
Batting training frequency	0.048	−0.113	0.187	−0.063
Batting training time (hour)	0.057	−0.098	0.186	−0.030

Abbreviation: IR, internal rotation; ER, external rotation; CON IR, concentric internal rotation; CON ER, concentric external rotation; ECC IR, eccentric internal rotation; ECC ER, eccentric external rotation; BW, body weight; ROM, range of motion; GIRD, glenohumeral internal rotation deficit; KJOC, Kerlan-Jobe Orthopedic Clinic Shoulder and Elbow score; U/E, upper extremity.

*p < 0.05, ** p < .001.

3.3 The results of stepwise multiple linear regression

Table 4 presents the results of stepwise regression analyses identifying significant predictors of CON and ECC IR and ER torques, based on variables identified through Pearson correlation analysis.

Table 4.
The results of stepwise multiple linear regression (n = 80).

Dependent variables Independent variables B SE β t(p) TOL VIF

CON IR torque CON ER torque 0.458 0.080 0.538 5.710 (<.001) 0.848 1.180

IR ROM 0.128 0.058 0.207 2.194 (0.031) 0.848 1.180

F(p) = 27.834 (<.001), R² = 0.420, adj. R² = 0.405, DW = 1.755

CON ER torque CON IR torque 0.589 0.089 0.501 6.647 (<.001) 0.796 1.256

ECC ER torque 0.551 0.085 0.438 6.450 (<.001) 0.979 1.021

GIRD −0.236 0.078 −0.229 −3.020 (0.003) 0.789 1.267

Experience 1.152 0.415 0.190 2.779 (0.007) 0.971 1.030

F(p) = 36.496 (<.001), R² = 0.661, adj. R² = 0.643, DW = 1.652

ECC IR torque ECC ER torque 0.366 0.078 0.455 4.672 (<0.001) 1.000 1.000

Height −0.217 0.084 −0.251 −2.580 (0.012) 1.000 1.000

F(p) = 14.165 (<.001), R² = 0.269, adj. R² = 0.250, DW = 2.407

ECC ER torque ECC IR torque 0.502 0.115 0.404 4.353 (<.001) 0.983 1.017

CON ER torque −0.303 0.074 0.381 4.100 (<.001) 0.983 1.017

F(p) = 20.559 (<.001), R² = 0.348, adj. R² = 0.331, DW = 1.783

Dependent variables	Independent variables	B	SE	β	t(p)	TOL	VIF
CON IR torque	CON ER torque	0.458	0.080	0.538	5.710 (<.001)	0.848	1.180
IR ROM	0.128	0.058	0.207	2.194 (0.031)	0.848	1.180
F(p) = 27.834 (<.001), R² = 0.420, adj. R² = 0.405, DW = 1.755
CON ER torque	CON IR torque	0.589	0.089	0.501	6.647 (<.001)	0.796	1.256
ECC ER torque	0.551	0.085	0.438	6.450 (<.001)	0.979	1.021
GIRD	−0.236	0.078	−0.229	−3.020 (0.003)	0.789	1.267
Experience	1.152	0.415	0.190	2.779 (0.007)	0.971	1.030
F(p) = 36.496 (<.001), R² = 0.661, adj. R² = 0.643, DW = 1.652
ECC IR torque	ECC ER torque	0.366	0.078	0.455	4.672 (<0.001)	1.000	1.000
Height	−0.217	0.084	−0.251	−2.580 (0.012)	1.000	1.000
F(p) = 14.165 (<.001), R² = 0.269, adj. R² = 0.250, DW = 2.407
ECC ER torque	ECC IR torque	0.502	0.115	0.404	4.353 (<.001)	0.983	1.017
CON ER torque	−0.303	0.074	0.381	4.100 (<.001)	0.983	1.017
F(p) = 20.559 (<.001), R² = 0.348, adj. R² = 0.331, DW = 1.783

Abbreviation: CON IR, concentric internal rotation; CON ER, concentric external rotation; ECC IR, eccentric internal rotation; ECC ER, eccentric external rotation; ROM, range of motion; GIRD, glenohumeral internal rotation deficit.

3.3.1 CON IR torque

The model for CON IR torque was significant (F = 27.834, p < .001), explaining 40.5% of variance (adj. R² = 0.405). CON ER torque (β = 0.538) and IR ROM (β = 0.207) were significant predictors, with CON ER torque exerting the stronger influence (Table 4).

3.3.2 CON ER torque

The regression model for CON ER torque was significant (F = 36.496, p < .001), explaining 64.3% of the variance (adj. R² = 0.643). Significant predictors included CON IR torque (β = 0.501), ECC ER torque (β = 0.438), GIRD (β = −0.229), and baseball experience (β = 0.190), with CON IR torque showing the strongest association (Table 4).

3.3.3 ECC IR torque

The model for ECC IR torque was significant (F = 14.165, p < .001), explaining 25.0% of variance (adj. R² = 0.250). ECC ER torque (β = 0.455) and height (β = −0.251) were significant predictors, with ECC ER torque demonstrating the greater influence (Table 4).

3.3.4 ECC ER torque

The regression model for ECC ER torque was significant (F = 20.559, p < .001), explaining 33.1% of the variance (adj. R² = 0.331). Significant predictors included ECC IR torque (β = 0.404) and CON ER torque (β = 0.381), with ECC IR torque showing the stronger influence (Table 4).

4 Discussion

This study revealed that adolescent baseball players with GIRD exhibited significantly lower CON IR and ER torques, along with an increased ECC IR/CON ER ratio, compared to those without GIRD. Concentric torque values had moderate associations with GH joint rotational ROM, particularly IR ROM and the presence of GIRD, which were identified as key influencing factors. Among these variables, CON ER torque demonstrated the strongest association with CON IR torque and was also correlated with baseball experience and GIRD. In contrast, eccentric torque values were not significantly associated with GH joint ROM. However, ECC IR torque was influenced by ECC ER torque and participant height, whereas ECC ER torque was affected by both ECC IR and CON ER torques. These findings imply that shoulder rotational ROM and the balance between agonist and antagonist muscle groups are critical determinants of torque production in adolescent baseball players.

The GIRD group exhibited concentric torque values of approximately 34.8%BW for IR and 35.83%BW for ER, which were about 5%BW and 7%BW lower, respectively, than those observed in the non-GIRD group. This indicates an insufficient IR ROM may affect the IR and ER muscle force generation, particularly during concentric contraction. A plausible explanation for this reduced concentric torque in the GIRD group is diminished muscle strength resulting from limited ROM.^23,46 Restricted IR ROM, i.e., 60.7° in the GIRD group compared to 75.1° in non-GIRD group, and a 6° reduction in total rotation may impaired optimal muscle length and joint mechanics, thereby reducing force generation in both IR and ER. These findings are supported by the observed poor-to-fair correlations between GH joint rotation ROM and concentric torque variables. Regression analysis also revealed a positive association between IR ROM and CON IR torque (β = 0.207), and a negative association between GIRD and CON ER torque (β = −0.229). These results support the hypothesis that limited ROM and shortened muscle length impair force production. Prior studies suggest that increased muscle length enhances energy absorption during eccentric contraction, thereby facilitating greater concentric force output.^47,48 Thus, reduced IR and total rotation ROM in the GIRD group may have limited eccentric energy absorption, resulting in decreased concentric torque in both IR and ER muscles. While these reductions in torque are consistent with the proposed mechanism, their relatively small magnitude requires cautious interpretation, as they may overlap with measurement precision. Hill et al.⁴² demonstrated ICC values between 0.73 and 0.93 in the supine position, supporting the reproducibility of this protocol and reinforcing the appropriateness of our measurement approach. However, Forthomme et al.⁴¹ reported SEM values ranging from 2.2 to 6.7 N·m and SDD values between 6.2 and 18.6 N·m across different testing positions, indicating that small torque differences may fall within the bounds of measurement error. The effect sizes calculated in our study (ES = 0.72 for IR and ES = 0.95 for ER) nevertheless suggest moderate-to-large magnitudes, and diminished rotator strength has been linked to increased injury risk.²⁰ Taken together, the torque differences observed in our study may have clinical implications related to performance and injury risk, but they should be interpreted with caution in light of measurement precision.

Eccentric overload of posterior shoulder soft tissues has been linked to inflammatory responses and morphological changes that reduce muscle strength.^7–9^,49 GIRD is associated with stiffness and inflexibility caused by repetitive overhead throwing, leading to connective tissue damage, inflammation, and external rotator weakness.^14,24,26 Guney et al.²⁶ reported lower eccentric ER torque (GIRD: 0.43 vs. non-GIRD: 0.75) and ECC ER/CON IR ratio in individuals with GIRD (GIRD: 0.55 vs. non-GIRD: 0.83), attributing this to a posterior-superior shift of the GH rotation axis. Previous studies have identified GIRD as a risk factor for throwing-related injuries with IR ROM differences of 18°¹⁵ and a fourfold increase in injury risk when GIRD exceeds 25°.¹⁹ Structural and strength alterations in rotator cuff muscles have also been documented in athletes with GIRD,^20,25 including reduced ER strength and decreased supraspinatus and infraspinatus thickness. Yamaura et al.²⁵ reported a higher isometric IR strength ratio and a lower ER strength ratio of the throwing shoulder compared to the non-throwing shoulder in professional baseball players with GIRD, and further demonstrated reduced supraspinatus and infraspinatus muscle thickness in the GIRD group. Tyler et al.²⁰ reported that 57% of overhead athletes with a history of injury had a GIRD of more than 20°, and athletes with moderate injuries exhibited weakened isometric supraspinatus strength. Despite these previous findings, the present study found no significant differences in ECC IR or ER torque between groups, and no association between eccentric torque and GH joint ROM. Interestingly, the ECC ER/CON IR ratio showed a tendency to be higher in the GIRD group, which contrasts with prior reports of decreased ratios. Guney et al.,²⁶ who examined adolescent volleyball and basketball athletes, found lower ECC ER torque and reduced ECC ER/CON IR ratios in the GIRD group (0.43 Nm) than the non-GRID group (0.75 Nm), but no significant differences in CON IR torque (GIRD: 0.81 Nm vs. non-GRID: 0.82 Nm). In contrast, our study did not observe differences in ECC ER torque between groups (GIRD: 41.6%BW vs. non-GIRD: 42.9%BW). Instead, we found a significant reduction in CON ER torque in the GIRD group. These distinct results likely contributed to the relatively higher ECC ER/CON IR ratio observed in our study. Regression analysis further indicated that ROM limitations may impair concentric torque generation, supporting the interpretation that reduced CON ER torque in the GIRD group was influenced by restricted mobility. Previous authors have suggested that eccentric overload of posterior shoulder soft tissues contributes to inflammatory responses and morphological changes that reduce muscle strength.^7–9^,49 Because our study included healthy adolescent baseball players without pain or prior injury, eccentric strength including ECC ER torque may not have been affected. This may explain why eccentric torque differences were not observed. Future studies should investigate how structural and pathological changes in posterior shoulder tissues influence eccentric torque generation. Methodological differences such as testing velocity (Our study: 120°/s vs. Guney et al.²⁶: 90°/s) and testing position (Our study: supine vs. Guney et al.²⁶: seated) may also have influenced torque outcomes. In our study, isokinetic torque was measured at a relatively higher angular velocity (120°/s), which, according to the force–velocity relationship, may have contributed to reduced concentric torque values, as concentric torque decreases with increasing velocity while eccentric torque tends to increase.^50,51 In addition, eccentric torque measurements are known to be less reproducible and more susceptible to the influence of non-contractile tissue properties.^52,53 Therefore, eccentric torque otucomes should be interpreted with caution, taking measurement reliability into consideration.

In this study, IR and ER torques demonstrated fair positive correlations with their respective antagonists in both concentric (r = 0.619) and eccentric (r = 0.454) contractions. Regression analysis revealed standardized coefficients of 0.538 for CON ER torque and 0.501 for CON IR torque, indicating mutual influence. For ECC IR torque, ECC ER torque exerted the greatest relative impact (β = 0.455), and similarly, ECC IR torque was the strongest predictor of ECC ER torque (β = 0.404). These findings imply that antagonist torque plays a significant role in agonist torque generation across both contraction types. In other words, higher antagonist torque was associated with greater agonist torque. Additionally, reciprocal influence was observed between concentric and eccentric ER torques. CON ER torque was the second strongest predictor of ECC ER torque (β = 0.381), while ECC ER torque was the second strongest predictor of CON ER torque (β = 0.438). These results suggest that strengthening antagonist muscles should be considered alongside IR and ER agonists when designing muscle training programs. Moreover, these results highlight the necessity of incorporating balanced training strategies that address both concentric and eccentric contractions for ER muscles. However, theoretical evidence explaining the mechanisms through which antagonist torque influences agonist torque remains limited. This relationship may not directly reflect physiological factors, such as muscle cross-sectional area, motor unit recruitment, firing rate, or energy metabolism. Although these are key determinants of muscle strength and power, their contributions were not confirmed in this study. Therefore, further research is warranted to elucidate the interactions between these physiological factors and torque generation. In addition to these strength-related findings, regression analysis emphasized the critical role of joint mobility in concentric torque generation. Specifically, IR ROM (β = 0.207) and the presence of GIRD (β = −0.229) were significant contributors to concentric torque production, indicating that mobility parameters complement muscular strength in determining shoulder performance. Accordingly, effective athlete management should encompass both the strengthening of agonist and antagonist muscles and targeted interventions to improve IR ROM, such as posterior shoulder stretching. Taken together, these strategies may help prevent shoulder dysfunction and support performance enhancement in adolescent baseball players with GIRD.

There are several limitations of this study. First, the physiological length characteristics of intramuscular soft tissue were not directly investigated. Examining correlations that incorporate additional factors influencing muscle strength, such as changes in muscle length at the musculotendinous junction, will be necessary to clarify further the relationship between GH joint ROM alterations and muscle torque. Second, we did not investigate concentric and eccentric torque in relation to other potential contributing factors, including risk factors for throwing-related injuries or scapular kinetics and kinematics. Therefore, future studies are needed to determine more clearly the causal relationship among the factors contributing to CON and ECC IR and ER muscle strength by accounting for these additional variables.

5 Conclusion

This study found that CON IR and ER torques were significantly reduced in athletes with GIRD compared to those without GIRD, and were closely associated with GH joint rotational ROM. IR ROM and the presence of GIRD emerged as key factors influencing concentric torque generation. Additionally, both concentric and eccentric torque production demonstrated strong interdependence between agonist and antagonist muscles, underscoring the role of the antagonist in supporting agonist strength. In contrast, ECC IR and ER torques did not differ significantly between groups and showed no meaningful correlation with GH joint ROM. These findings imply that evaluating IR ROM loss and GIRD, together with management strategies such as posterior shoulder stretching to improve IR ROM, may be essential for enhancing concentric IR and ER strength. Importantly, strengthen antagonist muscles within the same contraction type should be incorporated for both IR and ER, while the reciprocal influence observed between concentric and eccentric ER torques highlights the necessity of balanced training across both contraction modes. This integrated approach may provide a comprehensive strategy to optimize shoulder function in adolescent baseball players with GIRD.

Footnotes

Acknowledgements

We would like to thank everyone who contributed to this study.

ORCID iDs

Il-Young Yu

Na-Dan Lim

Ho-Kyung Choi

Tae-Gyu Kim

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was supported by the Global Joint Research Program funded by the Pukyong National University (202412370001).

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Certificate of English editing

We confirm that the English in this manuscript has been thoroughly edited by at least two professional editors, both of whom are native speakers of English. A certificate of English editing is available at:

References

Lyman

Fleisig

Andrews

. Effect of pitch type, pitch count, and pitching mechanics on risk of elbow and shoulder pain in youth baseball pitches. Am J Sports Med 2002; 30: 463–468.

Rechel

Yard

Comstock

. An epidemiologic comparison of high school sports injuries sustained practice and competition. J Athl Train 2008; 43: 197–204.

Saper

Pierpoint

Liu

, et al. Epidemiology of shoulder and elbow injuries among United States high school baseball players: school years 2005-2006 through 2014-2015. Am J Sports Med 2018; 46: 37–43.

Tajika

Oya

Kuboi

, et al. Risk factors for throwing-related shoulder and elbow pain in adolescent baseball players: a prospective study of physical and developmental factors. Orthop J Sports Med 2021; 9: 23259671211017129.

Fleisig

Bolt

Fortenbaugh

, et al. Biomechanical comparison of baseball pitching and long-toss: implications for training and rehabilitation. J Orthop Sports Phys Ther 2011; 41: 296–303.

Goes

Flores

Damer

, et al. Shoulder and elbow injuries in adult overhead throwers: imaging review. Radiographics 2023; 43: e230094.

Burkhart

Morgan

Kibler

. The disabled throwing shoulder: spectrum of pathology part I: pathoanatomy and biomechanics. Arthroscopy 2003; 19: 404–420.

Digiovine

Jobe

Pink

, et al. An electromyographic analysis of the upper extremity in pitching. J Shoulder Elbow Surg 1992; 1: 15–25.

Wilk

Meister

Andrews

. Current concepts in the rehabilitation of the overhead throwing athlete. Am J Sports Med 2002; 30: 136–151.

10.

Nakamizo

Nakamura

Nobuhara

, et al. Loss of glenohumeral internal rotation in little league pitchers: a biomechanical study. J Shoulder Elbow Surg 2008; 17: 795–801.

11.

Thacher

Varady

Khilnani

, et al. Current concepts on the management of shoulder instability in throwing athletes. Curr Rev Musculoskelet Med 2024; 17: 353–364.

12.

Borsa

Laudner

Sauers

. Mobility and stability adaptations in the shoulder of the overhead athlete: a theoretical and evidence-based perspective. Sports Med 2008; 38: 17–36.

13.

Burkhart

Morgan

Kibler

. The disabled throwing shoulder: spectrum of pathology. Part II: evaluation and treatment of SLAP lesions in throwers. Arthroscopy 2003; 19: 531–539.

14.

Kibler

Sciascia

Moore

. An acute throwing episode decreases shoulder internal rotation. Clin Orthop Relat Res 2012; 470: 1545–1551.

15.

Wilk

Macrina

Fleisig

, et al. Correlation of glenohumeral internal rotation deficit and total rotational motion to shoulder injuries in professional baseball pitchers. Am J Sports Med 2011; 39: 329–335.

16.

Aldridge

Stephen Guffey

Whitehead

, et al. The effects of a daily stretching protocol on passive glenohumeral internal rotation in overhead throwing collegiate athletes. Int J Sports Phys Ther 2012; 7: 365–371.

17.

Fairall

Cabell

Boergers

, et al. Acute effects of self-myofascial release and stretching in overhead athletes with GIRD. J Bodyw Mov Ther 2017; 21: 648–652.

18.

Myers

Laudner

Pasquale

, et al. Glenohumeral range of motion deficits and posterior shoulder tightness in throwers with pathologic internal impingement. Am J Sports Med 2006; 34: 385–391.

19.

Shanley

Rauh

Michener

, et al. Shoulder range of motion measures as risk factors for shoulder and elbow injuries in high school softball and baseball players. Am J Sports Med 2011; 39: 1997–2006.

20.

Tyler

Mullaney

Mirabella

, et al. Risk factors for shoulder and elbow injuries in high school baseball pitchers: the role of preseason strength and range of motion. Am J Sports Med 2014; 42: 1993–1999.

21.

Keller

De Giacomo

Neumann

, et al. Glenohumeral internal rotation deficit and risk of upper extremity injury in overhead athletes: a meta-analysis and systematic review. Sports Health 2018; 10: 125–132.

22.

Clarsen

Bahr

Andersson

, et al. Reduced glenohumeral rotation, external rotation weakness and scapular dyskinesis are risk factors for shoulder injuries among elite male handball players: a prospective cohort study. Br J Sports Med 2014; 48: 1327–1333.

23.

Neumann

. Kinesiology of the musculoskeletal system: foundations for rehabilitation. 3rd ed. St Louis, USA: Elsevier, 2017.

24.

Trakis

McHugh

Caracciolo

. Muscle strength and range of motion in adolescent pitchers with throwing-related pain: implications for injury prevention. Am J Sports Med 2008; 36: 2173–2178.

25.

Yamaura

Mifune

Inui

, et al. Relationship between glenohumeral internal rotation deficit and shoulder conditions in professional baseball pitchers. J Shoulder Elbow Surg 2021; 30: 2073–2081.

26.

Guney

Harput

Colakoglu

, et al. The effect of glenohumeral internal-rotation deficit on functional rotator-strength ratio in adolescent overhead athletes. J Sport Rehabil 2016; 25: 52–57.

27.

Fleisig

Andrews

Dillman

, et al. Kinetics of baseball pitching with implications about injury mechanisms. Am J Sports Med 1995; 23: 233–239.

28.

Seroyer

Nho

Bach

, et al. Shoulder pain in the overhead throwing athlete. Sports Health 2009; 1: 108–120.

29.

Bak

Magnusson

. Shoulder strength and range of motion in symptomatic and pain-free elite swimmers. Am J Sports Med 1997; 25: 454–459.

30.

Lam

. A study of antagonist/agonist isokinetic work ratios of shoulder rotators in men who play badminton. J Orthop Sports Phys Ther 2002; 32: 399–404.

31.

Andrade Mdos

Fleury

de Lira

, et al. Profile of isokinetic eccentric-to-concentric strength ratios of shoulder rotator muscles in elite female team handball players. J Sports Sci 2010; 28: 743–749.

32.

Noffal

. Isokinetic eccentric-to-concentric strength ratios of the shoulder rotator muscles in throwers and nonthrowers. Am J Sports Med 2003; 31: 537–541.

33.

Saccol

Gracitelli

da Silva

, et al. Shoulder functional ratio in elite junior tennis players. Phys Ther Sport 2010; 11: 8–11.

34.

Scoville

Arciero

Taylor

, et al. End range eccentric antagonist/concentric agonist strength ratios: a new perspective in shoulder strength assessment. J Orthop Sports Phys Ther 1997; 25: 203–207.

35.

Leahy

Florkiewicz

Shotwell

. Isokinetic dynamometry for external and internal rotation shoulder strength in youth athletes: a scoping review. Int J Sports Phys Ther 2024; 19: 1521–1531.

36.

Ellenbecker

Davies

. The application of isokinetics in testing and rehabilitation of the shoulder complex. J Athl Train 2000; 35: 338–350.

37.

Carvalho

Coelho

Silva

, et al. Assessment of reliability in isokinetic testing among adolescent basketball players. Medicina (Kaunas) 2011; 47: 446–452.

38.

Monti

Fernandez-Fernandez

. The Kerlan-Jobe orthopaedic clinic shoulder & elbow score used as a patient-reported outcome measure for the youth and high school aged baseball athlete. Int J Sports Phys Ther 2022; 17: 879–886.

39.

Morrissey

Driver

. Manual landmark identification and tracking during the medial rotation test of the shoulder: an accuracy study using three-dimensional ultrasound and motion analysis measures. Man Ther 2008; 13: 529–535.

40.

Wilk

Reinold

Macrina

. Glenohumeral internal rotation measurements differ depending on stabilization techniques. Sports Health 2009; 1: 131–136.

41.

Forthomme

Dvir

Crielaard

, et al. Isokinetic assessment of the shoulder rotators: a study of optimal test position. Clin Physiol Funct Imaging 2011; 31: 227–232.

42.

Hill

Pramanik

McGregor

. Isokinetic dynamometry in assessment of external and internal axial rotation strength of the shoulder: comparison of two positions. Isokinet Exerc Sci 2005; 13: 187–195.

43.

Kang

. The effects of posterior shoulder stretch on rotator cuff strength ratio in adolescent baseball players with scapular dyskinesis: a randomized controlled trial. Isokinet Exerc Sci 2018; 26: 63–71.

44.

Yildiz

Aydin

Sekir

, et al. Shoulder terminal range eccentric antagonist/concentric agonist strength ratios in overhead athletes. Scand J Med Sci Sports 2006; 16: 174–180.

45.

Carter

Kaminski

Douex

Jr , et al. Effects of high volume upper extremity plyometric training on throwing velocity and functional strength ratios of the shoulder rotators in collegiate baseball players. J Strength Cond Res 2007; 21: 208–215.

46.

Fowles

Sale

MacDougall

. Reduced strength after passive stretch of the human plantar flexors. J Appl Physiol 2000; 89: 1179–1188.

47.

Cavagna

. Elastic bounce of the body. J Appl Physiol 1970; 29: 279–282.

48.

Wilson

Wood

Elliott

. Optimal stiffness of series elastic component in a stretch-shorten cycle activity. J Appl Physiol 1991; 70: 825–833.

49.

Laudner

Moline

Meister

. The relationship between forward scapular posture and posterior shoulder tightness among baseball players. Am J Sports Med 2010; 38: 2106–2112.

50.

Duncan

Chandler

Cavanaugh

. Mode and speed specificity of eccentric and concentric exercise training. J Orthop Sports Phys Ther 1989; 11: 70–75.

51.

Walmsley

Pearson

Stymiest

. Eccentric wrist extensor contractions and the force velocity relationship in muscle. J Orthop Sports Phys Ther 1986; 8: 288–293.

52.

Kossev

Christova

. Discharge pattern of human motor units during dynamic concentric and eccentric contractions. Electroencephalogr Clin Neurophysiol 1998; 109: 245–255.

53.

Linnamo

Moritani

Nicol

, et al. Motor unit activation patterns during isometric, concentric and eccentric actions at different force levels. J Electromyogr Kinesiol 2003; 13: 93–101.

	Group
Variables	GIRD (n = 41)	Non-GIRD (n = 39)	Mean difference	95% CI	p
CON IR torque (%BW)	34.83 ± 6.37	39.44 ± 6.45	−4.61	−1.76, −7.46	0.002*
CON ER torque (%BW)	35.83 ± 7.66	42.71 ± 6.74	−6.88	−3.67, −10.10	<0.001*
ECC IR torque (%BW)	40.25 ± 5.66	41.11 ± 4.47	−0.86	−3.14, 1.41	0.452
ECC ER torque (%BW)	41.57 ± 5.87	42.88 ± 6.82	−1.31	−4.13, 1.52	0.360
ECC IR/CON ER ratio	1.17 ± 0.31	0.99 ± 0.19	0.19	0.07, 0.30	0.002*
ECC ER/CON IR ratio	1.23 ± 0.29	1.13 ± 0.31	0.10	−0.28, 0.24	0.121