Isokinetic shoulder rotator strength profiles in elite male Para athletes

Abstract

Background

Shoulder rotator strength balance is crucial for joint stability and injury prevention, particularly in Para athletes who rely heavily on upper limbs. However, isokinetic data across Para sports remain limited.

Methods

Fifty-three elite male Para athletes (throwers, sitting volleyball, Para swimming, and Para judo) underwent bilateral isokinetic testing of shoulder external (ER) and internal rotators (IR) at 60°, 180°, and 300°/s using a Biodex System 4 dynamometer. Peak torque relative to body weight (PT/BW), ER/IR ratios, and bilateral strength asymmetry (BSA) were calculated. Paired t-tests compared dominant and non-dominant sides within each sport, and a separate mixed-design ANOVA assessed dominance and sport effects.

Results

Dominant shoulders showed greater ER strength at 60°/s in para-throwers, sitting volleyball, and Para judo athletes (p < 0.05), while sitting volleyball players exhibited higher non-dominant ER strength at 300°/s (p < 0.001). Dominant IR strength was higher at 60°/s in sitting volleyball and Para judo, and at 300°/s in throwers (p < 0.05), whereas Para swimmers showed greater non-dominant IR strength at 300°/s (p = 0.046). Within-group analysis revealed significantly higher dominant-side ER/IR ratios in Para swimmers and Para judokas at 300°/s (p < 0.05). The mixed ANOVA confirmed significant dominance effects at 180°/s (p = 0.030, partial η² = 0.097) and 300°/s (p = 0.028, partial η² = 0.097), while group and interaction effects were non-significant. Inter-limb asymmetry remained ≤ 15% across sports.

Conclusions

Shoulder strength adaptations in elite Para athletes are primarily velocity-dependent rather than sport-specific. Functional asymmetries represent adaptive neuromuscular adjustments maintaining joint stability under asymmetric loads and provide reference data for training and injury prevention in Para sports.

Introduction

The rotator cuff muscles, particularly the internal and external rotators, play a crucial role in glenohumeral stability, overhead force generation, and the control of repetitive movements [1–3]. In Para athletes, these functions are especially critical, as limited lower-limb contribution or reduced trunk stability frequently shifts mechanical load to the shoulder complex [4–6]. This redistribution increases susceptibility to overuse, altered kinematics, and muscular imbalance, factors that collectively heighten the risk of chronic shoulder pain and injury across a wide range of Para sports [7–9]. Systematic assessment of shoulder strength can therefore inform performance monitoring in Paralympic contexts.

Effective glenohumeral stability depends on coordinated force transfer from the trunk and lower extremities; in athletes with physical impairments, reduced trunk or lower-limb contribution shifts mechanical demand proximally, increasing reliance on balanced rotator cuff force couples and dynamic stabilizers. Consequently, inter-limb asymmetry and agonist-antagonist balance may reflect both sport-specific neuromuscular adaptation and markers of mechanical overload and injury susceptibility. Isokinetic dynamometry, grounded in the torque-velocity relationship, enables velocity-specific quantification of concentric strength, bilateral deficits, and ER/IR ratios, providing a physiologically meaningful assessment of shoulder function beyond isometric or isotonic measures.

Isokinetic dynamometry provides valid and reproducible measures of muscle strength, agonist–antagonist balance, and inter-limb asymmetry, and is widely regarded as the gold standard for shoulder evaluation [10–12]. Despite its utility, research in Paralympic populations remains limited. Existing studies have largely focused on isolated impairment groups (e.g., spinal cord injury) or single sports, often at only one angular velocity [13–15].

These limitations constrain interpretation in three respects. First, the absence of multi-sport and multi-velocity data prevents differentiation between impairment-related and sport-specific neuromuscular adaptations. Second, the lack of comparative benchmarks limits clinicians’ ability to contextualize observed asymmetries or ER/IR ratios within performance or injury-risk frameworks. Third, the near absence of isokinetic data in high-demand Para sports, particularly Para judo, restricts the development of evidence-based conditioning strategies. This gap is particularly evident in Para swimming, where systematic evaluations of shoulder strength are scarce, and in Para judo, where no isokinetic data are available [16]. The lack of comparative evidence restricts both clinical translation and the development of sport-specific conditioning strategies. To address these gaps, this study provides a multi-sport isokinetic profile of shoulder internal and external rotators in elite male Paralympic athletes across three angular velocities. Different angular velocities (60°/s, 180°/s, and 300°/s) were selected to represent slow, moderate, and high contraction speeds, allowing the assessment of strength performance across a spectrum from maximal torque production to sport-specific high-velocity movements. By quantifying inter-limb asymmetry, ER/IR balance, and velocity-dependent responses, the study offers comparative data to inform performance monitoring and injury-risk management in Para sports.

Based on the identified gaps and study design, we hypothesized that: (1) dominant-side PT/BW would exceed non-dominant values across all angular velocities; (2) ER/IR ratios would differ between dominant and non-dominant limbs; (3) PT/BW and ER/IR ratios would differ between sports due to sport-specific mechanical demands; and (4) PT/BW would decrease with increasing angular velocity.

Materials and methods

Participants

The study involved 53 male elite Para athletes with a history of participation in the Paralympic Games, Asian Para Games, World Championships from Para athletics (throwing) (n = 18), sitting volleyball (n = 16), Para swimming (n = 8), and Para judo (n = 11). Although the athletes competed in different Para sports and presented heterogeneous impairment profiles, none had upper-limb impairments affecting shoulder function according to the Paralympic classification criteria. All participants met the functional eligibility requirements of their respective disciplines, demonstrating adequate trunk control, pelvic stability, and unrestricted upper-limb function relevant to testing procedures. For athletes with SCI, clinical evaluations confirmed the absence of neuropathological factors that could interfere with isokinetic performance. All SCI participants had lesions located between T10 and L1 levels, levels at which trunk control and abdominal muscle function are typically preserved [17, 18]. While sport-specific adaptations may exist, these shared functional prerequisites support meaningful comparison of shoulder rotator strength across disciplines.

Before the tests, participants were briefed on the purpose of the study, related procedures, and experimental risks, and all signed an informed consent document. None of the participants had experienced any shoulder-related issues within the past six months, including rotator cuff tears, shoulder surgery, or upper limb disabilities. Demographic information, including dominant hand, weekly training hours, and previous shoulder injuries, was recorded. All participants were free from any upper limb, trunk, or pelvic stability impairment that could physiologically or functionally influence shoulder strength or performance during the isokinetic assessment. Sample size was determined based on the availability of eligible Para athletes during the testing period, as comparable multi-sport isokinetic data in elite Para athletes were not available to inform an a priori effect size estimation. For transparency, a sensitivity analysis was conducted using G*Power 3.1, assuming a medium effect size (f = 0.25) based on Cohen’s conventional benchmarks for ANOVA designs (α = 0.05), which indicated an achieved power of 0.85 to detect moderate effects. The current research study was approved by the University of Tehran Ethics Committee (IR.UT. SPORT. REC.1400.043 – date of approval: December 20, 2021) and conducted in accordance with the principles set forth in the Helsinki Declaration. Table 1 provides a summary of the participants’ characteristics.

Table 1.

Participants’ characteristics

	N	Age (years)	Hight (cm)	Weight (kg)	Classification
		Mean ± SD	Mean ± SD	Mean ± SD
Para athletics (throwing)	18	31.83 ± 5.35	173.31 ± 22.11	110.71 ± 31.85	F11 (5), F12 (1), F13 (2), F34 (2), F41 (1), F55 (3), F57 (4)
Sitting volleyball	16	32.50 ± 9.27	189.19 ± 17.12	90.34 ± 22.83	VS1 (13), VS2 (3)
Para swimming	8	25.38 ± 3.96	180.50 ± 2.91	78.81 ± 2.93	S10 (7), S12 (1)
Para judo	11	30.73 ± 3.98	181.27 ± 8.41	94.59 ± 26.73	J2 (8), J1 (3)

cm centimeter, kg kilogram, SD standard deviation

Procedure

All isokinetic assessments were performed using a calibrated Biodex System 4 Pro isokinetic dynamometer (Biodex Medical Systems, Inc., Shirley, NY, USA). Shoulder external rotation (ER) and internal rotation (IR) were tested on both the dominant (D) and non-dominant (ND) sides, with dominance determined as the side used for throwing and writing [19]. All movements were executed in a concentric/concentric mode according to the standardized protocols described by Cools et al. and El-Ashker et al. [20, 21]. During testing, participants were seated and stabilized in the dynamometer chair using restraining straps across the chest and hips. The tested arm was corrected for gravity compensation, while the opposite arm grasped the side handle of the chair. Shoulder strength was measured with the arm positioned at 90° of abduction in the scapular plane, 20° of flexion in the frontal plane, and 90° of elbow flexion. The rotation range was set from 90° of ER to 0° of IR (90° total range of motion), and a range-of-motion stop ensured consistency across participants [14]. Prior to the actual test, participants performed a 5-min warm-up on an ergometer, followed by three to five submaximal familiarization trials to accustom them to the testing procedures [14]. The angular velocities of 60°/s, 180°/s, and 300°/s were selected in accordance with established isokinetic shoulder assessment protocols. These velocities reflect distinct aspects of neuromuscular performance: 60°/s primarily assesses maximal concentric strength under high torque conditions; 180°/s captures strength–power characteristics at moderate contraction speeds; and 300°/s evaluates high-velocity force production relevant to rapid and repetitive upper-limb actions [14, 20, 21]. The use of these standardized velocities facilitates comparison with previous studies in overhead and wheelchair-based athletes and allows examination of velocity-dependent adaptations in elite Para populations [14, 22]. The testing order was fixed (60°/s, 180°/s, 300°/s), progressing from slower to faster velocities in accordance with common isokinetic testing protocols to ensure movement familiarization and standardized assessment across participants.

Rest intervals of 1–2 minutes between sets and angular velocities were implemented in accordance with established isokinetic shoulder testing protocols to allow adequate phosphocreatine resynthesis and minimize fatigue during short-duration maximal contractions. Previous research in shoulder rotator assessments indicates that recovery periods of ≥ 60 s are sufficient to maintain peak torque performance across repeated efforts [23]. Six repetitions per velocity were selected based on widely adopted isokinetic procedures demonstrating that this range provides reliable peak torque measurement while limiting cumulative fatigue. Similar protocols have been employed in studies of overhead and wheelchair athletes, supporting methodological consistency and comparability. After a 5-min recovery, the contralateral shoulder was tested following the same protocol. To avoid systematic bias, the order of dominant and non-dominant shoulder testing was randomized. To ensure reliability and minimize measurement error, all assessments were conducted by the same experienced examiner familiar with the Biodex system. Throughout the procedures, participants received standardized verbal encouragement (e.g., “push as hard as possible” and “push as fast as possible”) to promote maximal effort. Testing was carried out during the preparatory training phase and outside the periods leading up to major competitions, thereby preventing any potential adverse effect on competitive performance. For subsequent analyses, peak torque relative to body weight (PT/BW) and ER/IR strength ratios were derived for both dominant and non-dominant extremities. The normalized PT/BW values correspond to the outputs automatically generated by the Biodex system software. Due to inter-sport variability in body mass, strength comparisons were conducted exclusively using normalized peak torque to ensure valid and comparable interpretations across disciplines. The bilateral strength asymmetry (BSA) equation was used to quantify interlimb difference. the direction of asymmetry between the D and ND limbs was determined individually for each participant. BSA is shown in Eq. 1:

1

Statistical analysis

The data were analyzed using Shapiro-Wilk tests to check for the normality of the distribution. Means and standard deviation (SD) were calculated for different parameters. Paired sample t-test was used to compare the PT/BW and ER/IR ratio between the dominant and non-dominant shoulders in the same participant. Separate 2-way mixed design ANOVA (with independent measures on groups and repeated measures on shoulder dominance) was used to assess differences in the relative ER/ IR strength ratios. The significance level was set at 0.05 and effect sizes were calculated using partial eta-squared (partial ɳ²), with 0.001, 0.06, and 0.14 classified as small, medium, and large effect sizes, respectively [24]. Effects sizes (Cohen’s d) were calculated to show practical differences between the two sides and were interpreted as: trivial (0–0.19), small (0.20–0.49), medium (0.50–0.79), and large (0.80 and greater) [25]. All analyses were performed using IBM SPSS Statistics for Windows, version 27.0 (IBM Corp., Armonk, NY, USA).

Results

In terms of ER, PT/BW muscle strength in the dominant shoulder was significantly higher than in the non-dominant shoulder for Para athletics (throwing), sitting volleyball, and Para judo players at an angular velocity of 60°/s (Para athletics (throwing), p = 0.029, d = 0.582; sitting volleyball, p = 0.008, d = 0.759; Para judo, p = 0.004, d = 1.117). No significant differences between the dominant and non-dominant shoulders were observed at 180°/s. At 300°/s, significantly higher ER strength was observed in the non-dominant shoulder of sitting volleyball players (p < 0.001, d = 1.030). However, in all sports except Para swimming, the non-dominant shoulder generally exhibited weaker strength than the dominant shoulder at all velocities, although these differences were not statistically significant (Table 2).

Table 2.

External and internal rotators’ strength of dominant and non-dominant shoulder

	External rotation		p-value	d	BSA	Internal rotation		p-value	d	BSA
	Dominant	Non-dominant				Dominant	Non-dominant
60°/s
Para athletics (throwing)	52.83 ± 9.55	48.74 ± 12.26	0.029*	0.582	13.41 (5.01, 21.80)	68.18 ± 17.72	64.93 ± 18.23	0.233	0.292	7.12 (-5.18, 19.43)
Sitting volleyball	51.68 ± 8.35	46.24 ± 6.85	0.008*	0.759	9.47 (2.10, 16.84)	64.98 ± 10.19	56.62 ± 12.04	0.002*	0.986	15.26 (7.13, 23.39)
Para swimming	49.40 ± 12.49	50.24 ± 10.08	0.839	0.074	-4.75 (-21.96, 12.47)	58.04 ± 11.24	65.06 ± 13.31	0.057	1.505	-12.83 (-26.08, 0.41)
Para judo	64.10 ± 5.27	56.15 ± 7.70	0.004*	1.117	12.28 (4.97, 19.60)	77.55 ± 16.90	68.65 ± 13.40	0.023*	0.809	9.85 (0.69, 19.01)
180°/s
Para athletics (throwing)	54.60 ± 11.10	49.88 ± 13.23	0.053	0.507	14.82 (3.55, 26.10)	63.53 ± 16.66	59.71 ± 17.48	0.149	0.356	9.33 (-3.65, 22.31)
Sitting volleyball	47.19 ± 6.34	45.99 ± 4.25	0.484	0.179	0.95 (-7.50, 9.39)	58.31 ± 11.21	53.02 ± 11.54	0.062	0.505	8.20 (-0.92, 17.31)
Para swimming	50.54 ± 7.97	48.21 ± 8.01	0.536	0.230	3.08 (-12.83, 19.00)	56.23 ± 9.43	63.01 ± 11.56	0.142	0.584	-13.46 (-32.41, 5.49)
Para judo	55.55 ± 4.01	52.85 ± 6.82	0.212	0.402	4.64 (-3.29, 12.57)	62.31 ± 12.09	63.85 ± 12.39	0.374	0.270	-2.91 (-9.70, 3.88)
300°/s
Para athletics (throwing)	51.84 ± 10.69	46.09 ± 13.79	0.062	0.487	13.67 (2.99, 24.35)	57.01 ± 14.19	53.41 ± 17.69	0.033*	0.568	3.07 (-9.83, 15.98)
Sitting volleyball	44.94 ± 6.48	50.82 ± 7.75	< 0.001*	1.030	-13.55 (-20.43, -6.67)	50.73 ± 11.11	53.11 ± 9.21	0.267	0.288	-7.87 (-17.47, 1.73)
Para swimming	53.90 ± 9.72	51.98 ± 11.39	0.667	0.159	2.09 (-14.89, 19.07)	52.43 ± 8.84	58.19 ± 6.91	0.046*	0.855	-12.51 (-24.79, -0.24)
Para judo	56.30 ± 8.00	51.45 ± 11.60	0.050	0.674	8.98 (0.71, 17.25)	56.03 ± 8.24	56.84 ± 10.02	0.705	0.118	-1.78 (-10.00, 6.44)

*Significant at p < 0.05

In terms of IR, PT/BW muscle strength in the dominant shoulder was significantly higher than in the non-dominant shoulder for sitting volleyball and Para judo athletes at an angular velocity of 60°/s (sitting volleyball, p = 0.002, d = 0.986; Para judo, p = 0.023, d = 0.809). No significant differences between dominant and non-dominant shoulders were observed at 180°/s. At 300°/s, significantly higher IR strength was observed in the dominant shoulder of Para athletics (throwing) (p = 0.033, d = 0.586), whereas at the same velocity, significantly higher IR strength was observed in the non-dominant shoulder of Para swimmers (p = 0.046, d = 0.855). In all sports, at all velocities, except Para swimming at 60°/s and 300°/s and Para judo at 180°/s and 300°/s, the non-dominant shoulder generally exhibited weaker strength than the dominant shoulder, but these differences were not statistically significant (Table 2).

Initially, within-group comparisons of the ER/IR ratio between the dominant (D) and non-dominant (ND) shoulders were performed using paired t-tests. No significant differences were observed in any group at 60°/s or 180°/s. However, at 300°/s, Para swimming and Para judo athletes exhibited a significantly higher ER/IR ratio in the dominant shoulder compared to the non-dominant shoulder (Para swimming, p = 0.028, d = 0.974; Para judo, p = 0.036, d = 0.729) (Table 3). No significant differences were found between the dominant and non-dominant shoulders in Para throwing or sitting volleyball athletes at any velocity.

Table 3.

ER/IR ratio in dominant and non-dominant shoulders across sport group

	Para athletics (throwing)		Sitting volleyball		Para swimming		Para judo		Dominance		Group		Dominance* Group
	D	ND	D	ND	D	ND	D	ND	P-value	Partial ƞ²	P-value	Partial ƞ²	P-value	Partial ƞ²
ER/IR 60°/s	78.16 ± 19.77	76.86 ± 18.94	79.16 ± 18.65	84.36 ± 17.06	84.68 ± 12.08	77.50 ± 7.99	88.30 ± 17.70	83.15 ± 11.82	0.649	0.005	0.544	0.046	0.357	0.069
ER/IR 180°/s	92.14 ± 23.74	84.36 ± 16.61	80.08 ± 12.25	86.21 ± 14.93	90.78 ± 12.65	77.56 ± 10.85	91.93 ± 16.82	84.38 ± 11.70	0.030*	0.097	0.456	0.053	0.083	0.131
ER/IR 300°/s	98.03 ± 23.95	88.46 ± 23.79	89.52 ± 17.21	96.87 ± 13.33	103.41 ± 14.91ᵞ	88.80 ± 12.31	101.86 ± 15.58 ᵞ	90.91 ± 13.96	0.028*	0.097	0.898	0.012	0.083	0.129

D dominant, ND non-dominant, Partial ƞ² partial eta-squared

*Significant at p < 0.05

Subsequently, a separate two-way mixed-design ANOVA was conducted with dominance (D vs. ND) as the within-subject factor and sport group (Para throwing, sitting volleyball, Para swimming, Para judo) as the between-subject factor. The dependent variable was the ER/IR ratio analyzed at three angular velocities (60°/s, 180°/s, and 300°/s). A significant main effect of dominance was found at 180°/s (p = 0.030, partial ɳ² = 0.096, medium effect) and 300°/s (p = 0.028, partial ɳ² = 0.097, medium effect), but not at 60°/s (p = 0.649, partial ɳ² = 0.005, small effect). No significant main effect of group was detected at any velocity (60°/s: p = 0.544; 180°/s: p = 0.456; 300°/s: p = 0.898; all small effects). Moreover, no significant interaction between dominance and group was observed (60°/s: p = 0.357, partial ɳ² = 0.069; 180°/s: p = 0.083, partial ɳ² = 0.131; 300°/s: p = 0.083, partial ɳ² = 0.129; all medium effects) (Table 3). Also, the interaction between shoulder dominance (dominant vs. non-dominant) and sport group for ER/IR strength ratios at three angular velocities (60°/s, 180°/s, and 300°/s) is illustrated in Figs. 1, 2 and 3.

Fig. 1.

Interaction between shoulder dominance and sport group for ER/IR strength ratios at 60°/s

Fig. 2.

Interaction between shoulder dominance and sport group for ER/IR strength ratios at 180°/s

Fig. 3.

Interaction between shoulder dominance and sport group for ER/IR strength ratios at 300°/s

Discussion

This study provides a comprehensive characterization of isokinetic shoulder rotator strength in elite male Para athletes across four upper-limb–dominant sports (Para athletics (throwing), Para swimming, Para judo, and sitting volleyball). Building upon limited data in Para populations, the present findings extend previous observations in overhead athletes without disabilities by demonstrating that shoulder strength adaptations are primarily influenced by side-specific and task-related demands rather than sport discipline alone [26–29]. However, differences in movement velocity, contraction type (cyclical vs. ballistic), unilateral versus bilateral loading, and cumulative shoulder stress may still contribute to subtle discipline-specific neuromuscular adaptations. Significant main effects of limb dominance were identified at specific angular velocities, whereas no group or interaction effects emerged. Overall, both inter- and intra-limb asymmetries remained within clinically acceptable limits, suggesting that functional balance and neuromuscular control may be preserved despite the repetitive, asymmetric loading inherent to upper-limb Para sports.

Although no statistically significant group effects were observed, discipline-specific movement patterns provide a plausible biomechanical context for the directional trends identified. Para throwers are exposed to repetitive unilateral, high-velocity rotational loading, which favors dominant-side internal and external rotator adaptations. In contrast, Para judo emphasizes sustained gripping, pulling, and isometric stabilization, potentially enhancing rotator co-contraction and joint stability rather than unilateral torque asymmetry. Sitting volleyball requires rapid overhead actions combined with continuous bilateral upper-limb support in a seated posture, demanding dynamic scapulothoracic control and symmetrical stabilization. Para swimming, characterized by cyclical bilateral propulsion, promotes more symmetrical torque production, particularly at higher velocities. These contrasting loading profiles may help contextualize the observed directional trends, even in the absence of statistically significant between-group differences.

The dominance of internal rotator (IR) strength observed in sitting volleyball and para judo athletes at 60°/s, and in Para throwers at 300°/s, may reflects the task-specific neuromuscular demands of striking, grappling, and throwing actions that rely heavily on powerful internal rotation. These findings are consistent with previous evidence in overhead athletes without disabilities demonstrating dominant-side IR superiority as a functional adaptation to repetitive loading [30–32]. However, unlike several reports in overhead throwers without disabilities demonstrating pronounced inter-limb asymmetries exceeding 15%, the present cohort exhibited more symmetrical torque profiles. This difference may reflect adaptive conditioning strategies, structured rehabilitation practices, or biomechanical constraints associated with impairment that modulate unilateral overload patterns in Para athletes.

Conversely, the greater non-dominant IR strength observed in Para swimmers at higher velocities (300°/s) likely reflects the symmetrical propulsion and bilateral stroke mechanics inherent to swimming. Although group effects were not statistically significant, the observed directional trends across sports may reflect discipline-specific loading characteristics [33, 34]. While enhanced IR capacity supports sport-specific performance, disproportionate dominance without corresponding ER strength may, in the long term, compromise glenohumeral stability and increase susceptibility to overuse injuries [35]. Accordingly, preventive conditioning emphasizing eccentric ER training and scapular stabilizing control remains essential to maintain shoulder health and dynamic balance.

External rotator (ER) strength was significantly greater in the dominant shoulder for Para throwers, sitting volleyball, and Para judo athletes at 60°/s, reflecting neuromuscular adaptations associated with the deceleration and control of repetitive high-velocity internal rotation. This dominant-side advantage aligns with previous findings in overhead Para athletes, where enhanced ER strength serves a stabilizing and protective function against anterior shoulder stress [30, 31, 33, 34]. In contrast, the greater non-dominant ER strength observed in sitting volleyball players at 300°/s may represent a compensatory adaptation facilitating bilateral stability and seated balance during rapid play sequences. Despite the absence of significant group effects, the velocity-dependent pattern of ER activation may indicate discipline-sensitive variations in concentric strength capacity, potentially contributing to shoulder stability and movement precision during high-speed rotational actions. Although the present isokinetic protocol assessed only concentric contractions, these findings have direct implications for the eccentric control phases inherent to overhead Para sport techniques. Inadequate eccentric resilience of the external rotators may compromise the shoulder’s ability to absorb rotational load, heightening the risk of cumulative microtrauma. Therefore, incorporating high-velocity concentric-eccentric ER exercises and scapular stabilization drills into individualized training programs is recommended to preserve dynamic glenohumeral stability, enhance performance efficiency, and mitigate overuse-related injury risk in elite Para athletes.

Across all sports, inter-limb asymmetry remained ≤ 15%, indicating balanced bilateral shoulder development despite the inherently asymmetric demands of most Para sports. This finding is clinically relevant, as previous research has shown that exceeding 10–15% asymmetry in overhead athletes, particularly in throwing and swimming populations, is associated with elevated injury risk and reduced joint efficiency [36]. The ability of elite Para athletes to maintain such functional symmetry likely reflects targeted strength and conditioning practices, as well as effective rehabilitation strategies emphasizing bilateral strength retention. Rather than indicating perfect equality, this symmetry represents a functional equilibrium that allows for performance optimization while preserving shoulder integrity under repetitive, unilateral loading conditions. Additionally, Para athletes’ specific features such as altered trunk contribution, compensatory upper-limb reliance due to lower-limb impairment, and long-term adaptation to assistive device use may influence shoulder torque production patterns. These unique biomechanical contexts differentiate Para athletes from athletes without disabilities counterparts and may partly explain the relatively preserved bilateral symmetry observed in the present sample.

ER/IR ratio generally ranged between 0.7 and 1.0 across all sports, falling within the physiological range reported for healthy shoulders and suggesting adequate coordination between agonist and antagonist rotator groups. The ratio differed significantly between dominant and non-dominant shoulders at 180°/s and 300°/s, demonstrating a velocity-dependent modulation of shoulder rotator balance. Medium effect sizes indicate that these differences are functionally meaningful, even in the absence of significant group or interaction effects. The Ratios falling within the physiological range reported for healthy shoulders and suggesting adequate coordination between agonist and antagonist rotator groups. The increase in ratio values at higher angular velocities reflects enhanced external rotator efficiency in deceleration control and dynamic joint stabilization during rapid movements. These velocity-dependent differences may be underpinned by neural adaptations, including altered motor unit recruitment strategies, enhanced rate-of-force development, and improved intermuscular coordination developed through repetitive high-speed training exposure. These adaptations likely represent a protective mechanism developed through repetitive high-velocity training, enabling Para athletes to maintain joint integrity under asymmetric loading. The absence of inter-sport variation may indicate shared neuromuscular characteristics among elite Para athletes, potentially influenced by similar training and rehabilitation approaches. Ongoing monitoring of the ER/IR ratio and targeted eccentric ER strengthening remain essential for sustaining shoulder balance, optimizing performance, and minimizing overuse-related injury risk in overhead Para sports.

Interestingly, the lack of significant dominance effects at lower testing speeds (60°/s) further supports the interpretation that shoulder strength asymmetries in Para athletes are primarily functional rather than structural. At submaximal velocities, the rotator cuff muscles operate within stable torque ranges that minimize inter-limb performance gaps. However, as testing speed increases to 180°/s and 300°/s, velocity-dependent demands amplify neuromuscular differentiation between dominant and non-dominant shoulders, revealing the adaptive specialization of dynamic stabilizers under sport-specific loading. This finding reinforces the notion that high-velocity conditions better expose performance-relevant asymmetries and should therefore be prioritized in both assessment and preventive conditioning protocols.

At the statistical level, both paired-sample and mixed-design analyses consistently indicated that side-to-side differences in the ER/IR ratio were velocity-dependent rather than sport-specific. The mixed ANOVA revealed significant main effects of dominance at 180°/s and 300°/s, suggesting that overall neuromuscular balance in Para athletes is modulated by task velocity and differential activation of stabilizing musculature under high rotational demand. Complementary paired-sample comparisons further demonstrated that this dominance effect at 300°/s was primarily driven by Para swimmers and Para judokas, implying that the pooled asymmetry across the sample reflects disproportionate contributions from these disciplines. Although the dominance × group interaction narrowly missed statistical significance, the medium effect size indicates a trend toward sport-specific modulation that warrants verification in larger cohorts. These findings highlight the importance of targeted high-velocity eccentric ER conditioning and ongoing isokinetic monitoring, particularly for swimmers and judokas, whose repetitive high-speed internal rotation and grip-intensive techniques impose substantial deceleration demands. Collectively, these outcomes suggest that shoulder strength asymmetries in elite Para athletes appear to be more velocity-sensitive than discipline-dependent within the present sample.

Limitations

This study has several limitations that should be acknowledged. First, the cross-sectional design precludes causal inferences and limits the interpretation of temporal adaptations. Second, eccentric strength was not assessed, preventing the calculation of functional ER/IR ratios and limiting comprehensive evaluation of muscle balance. Third, the sample included only male elite Para athletes, which restricts generalizability to female athletes. Finally, sample size was determined by the availability of eligible national-level athletes during the testing period, which may limit statistical power for detecting small effects.

Conclusion

This study indicates that elite male Para athletes demonstrate velocity-sensitive, side-dominant yet functionally balanced shoulder rotator strength profiles across upper-limb–dominant Para sports. While significant dominance effects were observed at higher angular velocities, minimal group and interaction effects suggest that torque adaptations are influenced more by velocity-dependent neuromuscular demands than by sport category alone within the present sample.

ER/IR ratios and inter-limb asymmetries remained within established physiological thresholds, consistent with previous findings in overhead athletes. However, the relatively preserved bilateral symmetry observed here may also reflect Para sports specific biomechanical contexts, including compensatory upper-limb reliance and altered trunk contribution, which shape torque production patterns differently from athletes without disabilities populations.

The velocity-dependent modulation of rotator strength likely reflects neural and mechanical adaptations associated with repetitive high-speed training exposure. These findings underscore the importance of high-velocity assessment when identifying functional asymmetries relevant to performance and injury prevention.

Based directly on the observed directional torque trends, practical implications include grip–pull and deceleration-focused strengthening for Para judo athletes, bilateral propulsion symmetry training for Para swimmers, and upper-limb support–balance drills for sitting volleyball players. These recommendations are grounded in the velocity-dependent dominance patterns identified in this study rather than generalized conditioning principles.

Abbreviations

ER

External rotation

IR

Internal rotation

D

Dominant

ND

Non-dominant

PT/BW

Peak torque relative to body weight

BSA

Bilateral strength asymmetry

SD

Standard deviation

Authors’ contributions

All authors read and approved the final version of the manuscript. MM: Writing—original draft, Conceptualization, Methodology, Validation, Writing—review & editing, Data curation, Supervision, Investigation; FA: Writing—original draft, Data curation, Formal analysis, Software, Writing—review & editing; EKH: Writing—original draft, Data curation, Investigation; SH: Writing—review & editing, Data curation, Investigation, Software; NA: Writing—review & editing, Data curation, Investigation; SHM: Writing—review & editing, Data curation, Investigation; RS: Writing—review & editing, Data curation, Investigation.

Funding

No financial support was received for the conduct of this study or preparation of this manuscript.

Data availability

The dataset supporting the findings from this study is available from the corresponding authors upon request.

Declarations

Ethics approval and consent to participate

The current research study was approved by the University of Tehran Ethics Committee (IR.UT. SPORT. REC.1400.043 – date of approval: December 20, 2021) and conducted in accordance with the principles set forth in the Helsinki Declaration. Before the tests, participants were briefed on the purpose of the study, related procedures, and experimental risks, and all signed an informed consent document.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.