Abstract
Background:
Repetitive motion can alter joint angles and subsequently affect the control of the center of mass (CoM). While the CoM has been studied as a fatigue indicator in various sports, the control of the whole-body CoM during repetitive pitching in baseball pitchers has not been examined. This study aimed to investigate changes in lower-extremity joint angles and CoM control in collegiate baseball pitchers after repetitive pitching.
Hypothesis:
Baseball pitchers would exhibit significant increase in lower-extremity flexion angles, CoM position, and CoM variability after repetitive pitching.
Study Design:
Descriptive laboratory study.
Level of Evidence:
Level 3.
Methods:
A total of 23 pitchers from the Collegiate Baseball League were recruited. A motion analysis system was employed to assess lower-extremity joint angles and CoM position during the simulated game, while pitching accuracy and velocity were also recorded.
Results:
The results revealed a significant forward and downward shift in CoM position (P < 0.05), along with increased CoM variability in all directions (P < 0.05) after the simulated game. Furthermore, there was a significant increase in flexion angles of the knee and hip (P < 0.05); however, pitching velocity and accuracy did not demonstrate significant changes.
Conclusion:
Repetitive pitching leads to kinematic changes that should be monitored to prevent sports injuries.
Clinical Relevance:
Baseball pitchers have the ability to modify the control of their CoM and angles of their lower-extremity joints to sustain their pitching performance. It is crucial to monitor compensatory strategies closely to avoid shoulder and elbow injuries among these pitchers.
Repetitive pitching during baseball games can lead to fatigue and sports injuries in baseball pitchers.18,42 Epidemiological studies have found that 82% of baseball pitchers report arm fatigue during games and practice.2,26 Moreover, fatigue and playing position (pitchers or catchers) are risk factors for arm injuries in baseball players.28,36 Previous studies have demonstrated that either pitching accuracy or velocity tends to decrease after approximately 100 repetitive pitches.14,16 It is plausible that the pitchers might alter their body kinematics to compensate for such negative impacts on their performance. For instance, the angle of shoulder horizontal abduction increased at maximal shoulder external rotation, which in turn increased the anterior shoulder force and medial elbow force. 14 Such a change could increase valgus torque in the elbow and lead to ulnar collateral ligament strain.4,14 In fact, repetitive pitching has been demonstrated to cause upper-extremity injuries in baseball pitchers due to the compensation pattern observed in response. 8 To enhance the implementation of injury prevention strategies in baseball pitchers, it is crucial to examine the changes in kinematic patterns that occur as a result of compensatory movements caused by repetitive pitching.
Such compensatory patterns are not limited to the changes in the upper-extremity. As the pitching movement requires intensive involvement from the lower extremity, repetitive pitching might also increase the risk of lower-extremity injuries in players.17,37 In fact, the lower extremity is essential given its role in energy generation for pitching.25,31 Energy is generated in a particular order, flowing from the lower extremities through the trunk to the throwing arm during the pitching motion. Inadequate energy from the lower extremities may lead to compensatory actions by the trunk or upper extremities, thereby increasing the risk of upper-extremity injuries. Therefore, examining the changes in lower-extremity kinematics after repetitive pitching becomes necessary.
The compensatory pattern induced by repetitive movements alters joint angles, which in turn affects the control of the body’s center of mass (CoM).19,20 Numerous studies have investigated CoM position and variability after repetitive movements in a number of sports, such as basketball, soccer, and ultramarathons.24,38,41 The studies demonstrated that lower-extremity joint angles, position of the CoM, and variability of the CoM increased after repetitive movements. This could potentially have a negative impact on athletic performance.12,24,41 Taken together, control of the CoM was regarded as an important parameter related to stability and performance in response to repetitive movements. 35
Uncertainty still exists over whether lower-extremity joint angles, CoM position, and CoM variability increase as a result of repetitive pitching in baseball pitchers (Figure 1). The changes in lower-extremity joint angles, CoM position, and CoM variability during baseball games could be monitored carefully to minimize the risk of shoulder and elbow injuries due to overuse in baseball pitchers. Therefore, the aim of this study was to assess the lower-extremity joint angles and CoM control in collegiate baseball pitchers following a simulated game. It was hypothesized that the baseball pitchers would exhibit significant increases in lower-extremity joint flexion angles, change of CoM position, and CoM variability after repetitive pitching.

Conceptual framework and hypothesis for repetitive pitching-induced change of CoM and baseball pitching injury. CoM, center of mass.
Methods
Participants
Baseball pitchers were recruited from the University Baseball League (Open Division, Level 2). This league comprises Open Division Level 1, Level 2, and General Division. This study was approved by the research ethics committee of the university hospital. Participants who met the following inclusion criteria were considered eligible for the study: (1) age between 18 and 25 years; (2) pitchers’ experiences >5 years; (3) asymptomatic collegiate male baseball pitchers. Exclusion criteria were as follows: (1) history of shoulder or elbow surgery; (2) injury to the shoulder or elbow within 3 months before this study; (3) brachial plexus injury.
Experimental Setup
The pitching test was conducted in a 3-dimensional (3-D) motion analysis laboratory for a simulated game of 4 innings, with each inning consisting of 20 pitches (ie, a total of 80 repetitive pitches). We chose 80 pitches based on the literature and in consultation with coaches to avoid injury during the experiment.1,30 There was a 5-minute rest between each inning. During the pitching session, the pitcher threw toward the strike zone net located 18.4 meters away from a wooden portable mound. Each participant was instructed to pitch fastball with maximum effort. Pitching kinematics variables were determined by the conventional VICON Plug-in-Gait model.21,43 Static calibration was performed with 47 reflective markers; 4 medial markers were removed during the dynamic trials. 6 All markers were attached to the surface of the skin by a licensed physical therapist. Pitchers were told to warm up as they would before a really competitive game, with the limitation of pitching no more than 20 pitches before the start of the session and 10 pitches between the innings. Pitchers needed to pitch the ball within 20 seconds following the pitch clock rule. They were permitted to simulate real-game situations by using their regular pregame and interinnings routines. 27 Video data of pitching kinematics were recorded using a VICON motion analysis system with 7 cameras (Vicon MX 13 & 40, Oxford Metrics Ltd) at 240 Hz - the highest frequency of this system. 11 The position data of each marker was filtered using a second-order Butterworth double-pass filter, with a cutoff frequency of 13.4 Hz. 10
Pitching Motion
The pitching motion was divided into 4 events23,40: (1) stride foot contact, defined as the moment when the height of the foot marker attached to the stride leg was <0.06 cm above the pitching mound; (2) maximum shoulder external rotation (MER), an instance of maximum shoulder external rotation of the pitching arm; (3) ball release, defined as the instance when the distance between the ball and the hand marker on the pitching arm exceeds the mean between each marker calculated from the stride foot contact to MER; (4) maximum shoulder internal rotation (MIR) - an instance of maximum shoulder internal rotation of the pitching arm. The joint angles during the pitching motion were calculated using the VICON Nexus Version 2.5.0 (Oxford Metrics Ltd). All the events were calculated and determined using customized MATLAB R2010a software (MathWorks).
According to previous literature, this study selected the first 3 trials and the last 3 trials out of 80 trials to define the “before” and “after” of the simulated game. 32 VICON Nexus Version 2.5.0 software was utilized to calculate the lower-extremity joint angles (hip and knee joint angles). The CoM position and variability (Figure 2) of the first and last 3 trials were calculated using our customized MATLAB R2010a software. 6 Variability was the mean squared deviation from the mean of the 3 trials. The general formula used to calculate variability is
where Xi is the position of each trial and μ is the mean value of the position of trials.

CoM position variability was calculated for 4 events: stride foot contact (circles), MER (squares), ball release (triangles), and MIR (stars) using the first 3 (01, 02, 03) and the last 3 (78, 79, 80) trials of a representative participant. After, after a simulated game; Before, before a simulated game; CoM, center of mass; MER, maximum shoulder external rotation; MIR, maximum shoulder internal rotation.
Rating of Perceived Exertion
At the end of each inning, each pitcher reported their feeling of fatigue level using 15-points of Borg’s rating of perceived exertion (RPE) scale, with 6 indicating no fatigue and 20 being an extreme level of fatigue.7,13 The mean scores of RPE in the first inning and fourth inning were analyzed.
Pitching Accuracy and Velocity
If the baseball was in the strike zone, pitching accuracy was recorded as “strike”; however, if the baseball was outside the strike zone, pitching accuracy was recorded as “ball” for each pitching trial. Pitching velocity was determined for each attempt using a velocity radar gun (Stalker Sport, Applied Concepts Inc) for each trial. The mean values of the pitching accuracy and velocity of the first and the last 3 trials were evaluated.
Statistical Analysis
In this study, descriptive data of the participants were analyzed and presented as means and standard deviations for continuous variables. The normal distribution assumption of the data was assessed using the Shapiro-Wilk test. Since the data did not follow a normal distribution, nonparametric analysis was conducted. Specifically, the means of the first and last 3 trials in each pitching event were compared using the Wilcoxon signed-rank test. This analysis aimed to examine the differences in lower-extremity joint angles, CoM position, CoM variability, RPE, and pitching accuracy and velocity. Statistical significance was determined using a P value threshold of <0.05, indicating that differences with P values below this threshold were considered statistically significant. The statistical analysis was conducted using PASW Statistics Version 18 for Windows software (SPSS). By employing these statistical methods, the study aimed to provide an objective evaluation of the observed differences in the measured variables between the pre- and postsimulated game conditions.
Results
A total of 23 collegiate baseball pitchers were recruited. The pitchers were (mean ± SD) 20.0 ± 1.7 years old, 179.2 ± 5.1 cm tall, and had a mass of 78.0 ± 11.9 kg. A total of 17 pitchers were right-handed, while the other 6 pitchers were left-handed.
CoM Position and Lower-Extremity Joint Angles
Figure 3 illustrates the CoM position and Table 1 shows hip angle and knee angle at 4 pitching events.

CoM position at 4 pitching events. After, after a simulated game; Before, before a simulated game; MER, maximum shoulder external rotation; MIR, maximum shoulder internal rotation; X, forward (+) backward (-); Y, pitching arm side (+) nonpitching arm side (-); Z, upward (+) downward (-).
Hip and knee joint angles (°) at 4 pitching events before and after a simulated game
MER, maximum shoulder external rotation; MIR, maximum shoulder internal rotation.
Statistically significant, P < 0.05.
During stride foot contact event, the CoM position shifted significantly more forward (before, mean = 38.69 ± 18.30 cm; after, mean = 52.31 ± 30.12 cm; P = 0.03) and downward (before: mean = 18.38 ± 5.5 cm; after: mean = 26.75 ± 5.62 cm; P < 0.01) in the last 3 trials compared with the first 3 trials. The angles of hip flexion (before, mean = 66.92° ± 12.94°; after, mean = 76.12° ± 17.63°; P < 0.01) and knee flexion (before, mean = 41.35° ± 8.99°; after, mean = 88.27° ± 15.44°; P < 0.01) were significantly increased after a simulated game. The angle of hip internal rotation (before, mean = 7.17° ± 10.24°; after, mean = 1.44° ± 12.29°; P < 0.01) was also significantly increased in the last 3 trials.
During the MER event, the CoM position shifted significantly forward (before, mean = 73.99 ± 8.64 cm; after, mean = 101.20 ± 14.40 cm; P < 0.01) and downward (before, mean = 24.18 ± 5.17 cm; after, mean = 36.1 ± 7.17 cm; P < 0.01) in the last 3 trials compared with the first 3 trials. The angles of hip flexion (before, mean = 80.79° ± 11.69°; after, mean = 88.27° ± 15.44°; P < 0.01) and knee flexion (before, mean = 47.41° ± 10.92°; after, mean = 55.22° ± 8.36°; P = 0.03) were increased significantly after a simulated game. The angle of hip internal rotation (before, mean = 14.45° ± 5.24°; after, mean = 8.58° ± 7.55°; P = 0.02) was significantly increased after a simulated game.
During the ball release event, the CoM position shifted significantly to more forward (before, mean = 76.16 ± 8.33 cm; after, mean = 103.6 ± 15.01 cm; P < 0.01) and downward (before, mean = 24.60 ± 5.23 cm; after, mean = 36.29 ± 7.35 cm; P < 0.01) in the last 3 trials compared with the first 3 trials. The angles of hip flexion (before, mean = 80.24° ± 11.51°; after, mean = 94.27° ± 15.53°; P < 0.01) and knee flexion (before, mean = 44.16° ± 7.75°; after, mean = 51.08° ± 9.91°; P = 0.02) were increased significantly after a simulated game.
During the MIR event, the CoM position shifted significantly to more forward (before, mean = 89.4 ± 8.74 cm; after, mean = 121.6 ± 15.5 cm; P < 0.01) and downward (before, mean = 17.53 ± 5.44 cm; after, mean = 27.22 ± 8.31 cm; P < 0.01) in the last 3 trials compared with the first 3 trials. The angles of hip flexion (before, mean = 73.38° ± 12.39°; after, mean = 84.84° ± 16.02°; P < 0.01) and knee flexion (before, mean = 16.97° ± 4.71°; after, mean = 22.19° ± 5.68°; P < 0.01) were increased significantly after a simulated game. The angles of hip adduction (before, mean = 5.34° ± 5.14°; after, mean = 9.83° ± 5.22°; P < 0.01) and internal rotation (before, mean = 13.70° ± 5.71°; after, mean = 9.33° ± 8.43°; P = 0.04) were also increased significantly after a simulated game.
CoM Variability
The results of CoM variability were increased significantly during the last 3 trials at 4 pitching events after a simulated game (P < 0.05). The detailed results are shown in Figure 4.

CoM variability at 4 pitching events. After, after a simulated game; Before, before a simulated game; MER, maximum shoulder external rotation; MIR, maximum shoulder internal rotation; X, forward (+) backward (-); Y, pitching arm side (+) nonpitching arm side (-); Z, upward (+) downward (-). *Statistically significant, P < 0.05.
During stride foot contact event, the CoM variability in forward-backward (before, mean = 79.39 ± 110.62 cm2; after, mean = 198.81 ± 173.64 cm2; P < 0.01), pitching arm side-nonpitching arm side (before, mean = 48.89 ± 75.87 cm2; after, mean = 93.42 ± 91.54 cm2; P = 0.02) and upward-downward direction (before, mean = 32.77 ± 34.61 cm2; after, mean = 145.65 ± 74.62 cm2; P < 0.01) were increased significantly after a simulated game.
During the MER event, the CoM variability in forward-backward (before, mean = 16.63 ± 25.44 cm2; after, mean = 19.47 ± 3.99 cm2; P = 0.02) and upward-downward direction (before, mean = 15.47 ± 12.29 cm2; after, mean = 21.79 ± 8.29 cm2; P = 0.01) were increased significantly after a simulated game.
During the ball release event, the CoM variability in forward-backward (before, mean = 12.44 ± 10.95 cm2; after, mean = 20.44 ± 3.75 cm2; P < 0.01), pitching arm side-nonpitching arm side (before, mean = 27.73 ± 55.94 cm2; after, mean = 54.02 ± 60.33 cm2; P < 0.01) and upward-downward direction (before, mean = 16.10 ± 7.95 cm2; after, mean = 22.72 ± 6.88 cm2; P < 0.01) were increased significantly after a simulated game.
During the MIR event, the CoM variability in forward-backward (before, mean = 13.36 ± 7.40 cm2; after, mean = 27.11 ± 6.23 cm2; P < 0.01) was increased significantly after a simulated game.
Rating of Perceived Exertion
The results of RPE showed a significant increase (P < 0.01) during the fourth inning (mean = 15.7 ± 2.36) compared with the first inning (mean = 11.91 ± 2.19).
Pitching Accuracy and Velocity
Pitching accuracy was assessed by calculating the mean number of strikes in the first and last 3 trials (before, mean = 1.2 ± 1.0 strike number; after, mean = 1.3 ± 0.9 strike number; P = 0.39) and velocity was also measured in the first and last 3 trials (before, mean = 34.33 ± 2.47 m/s; after, mean = 34.44 ± 2.53 m/s; P = 0.16) were not significantly different between before and after the simulated game.
Discussion
The purpose of this study was to examine the changes in the lower-extremity joint angles and CoM control in collegiate baseball pitchers after a simulated game. The results support our hypothesis that lower-extremity joint angles, CoM position, and CoM variability were increased significantly after the simulated game. Baseball pitchers could alter the control of lower-extremity joint angles and CoM to maintain their pitching performance. As a result, decreases in pitching velocity or pitching accuracy cannot be solely relied upon as indicators for player substitution and injury prevention. The changes in the control of lower-extremity joint angles and CoM occurred before decreased velocity and accuracy. Therefore, monitoring and assessing changes in lower-extremity joint angles and CoM control could be useful indicators for player substitution and for implementing effective injury prevention strategies.
The lower-extremity joint angles increased significantly after the simulated baseball game. Hip flexion, external rotation, and knee flexion angles in the last 3 trials were all greater than those in the first 3 trials. Hip range of motion plays an important role in baseball pitching. However, no previous study has reported the changes in hip joint angles during a simulated game.5,22,44 Pitching motion was achieved by activating the kinetic chain from the lead leg, hip, and trunk to the pitching arm. 3 These body segments were coordinated during the pitching to generate and transfer energy.31,33 The hip joint was the primary joint that initiated trunk rotation. Insufficient hip range of motion has been shown to reduce the efficiency of force transfer from the lead leg to the pitching arm.22,44 An increased hip motion after repetitive pitching might be a strategy to optimize the kinetic chain.
Studies have shown in college or high school baseball players that the knee flexion angles increased as the number of pitching trials increased, which was consistent with our findings.7,9 However, other studies reported that knee flexion angles in professional baseball players decreased or remained unchanged after the simulated game.10,29 The inconsistent findings might result from the different experience levels of the study participants. Professional baseball players might have developed better fatigue resistance to manage their force consumption and maintain excellent pitching performance than college baseball players during repetitive pitching.
In our study, hip and knee angles increased after a prolonged pitching activity. Previous studies mentioned that increased joint angles led to increased force generation.2,34 However, the soft tissue elongated and compressed at the position with a large joint angle, which could lead to lower-extremity soft tissue injury. Therefore, pitchers should be aware of how their hip and knee angles alter over prolonged pitching to prevent injury.
The CoM shifted forward and downward after the simulated game, along with increases in the hip and knee flexion angles of the stride leg in our study. The aforementioned changes in these angles could be the reason why the CoM was shifted forward and downward. In previous studies, repetitive pitching has been shown to lead to a significant increase in knee flexion at ball release.7,29 These changes in joint angles may explain the observed forward and downward shift of the CoM, aligning with previous studies that demonstrated a significant increase in knee flexion during ball release as a result of repetitive pitching.
The results showed a significant increase in CoM variability after the simulated game. The variability of CoM position has been considered an important parameter related to the stability of performance. 35 After repetitive movements, greater postural sway can lead to an increase in the variability of CoM position. 39 In our results, CoM variability increased in all directions during each pitching event, and the angles of hip and knee flexion were also increased after a simulated game. The increased CoM variability in the forward-backward and upward-downward direction could be a result of increased hip and knee flexion angles. The increased joint angles might result from the decreased muscle force of the hip and knee extensors. In addition, the increased CoM variability in the left-right direction could be the result of decreased hip rotation angle after the simulated game. The decreased hip internal rotation might result from the fatigue state of the hip internal rotators. It would result in more postural sway, which in turn would result in an increase in CoM variability.
Consistent with a previous study, the pitching accuracy and velocity were not decreased significantly in our study after a simulated game. 16 However, other studies mentioned that pitching velocity decreased throughout the simulated game.7,15 The main difference between our study and other studies was the number of repetitive pitches. The total number of pitches ranged from 90 to 100 pitches in other studies,7,15 and was 80 pitches in our study. We chose 80 pitches based on the literature and in consultation with coaches to avoid injury during the experiment.1,30 Although pitching performance did not differ significantly after a simulated game in our study, we still found that lower-extremity joint angles and CoM variability were altered significantly. Baseball pitchers could alter the control of joint coordination and CoM before the decline in pitching performance.
Study Limitations
This study has the following limitations. First, the study was conducted in an indoor motion analysis laboratory, which may limit the similarity of the simulated game to a real match-playing environment. Second, the participants were recruited only from universities. Therefore, the results do not reflect all levels of baseball pitchers. Third, the study did not include a sample size calculation, potentially leading to statistical errors related to sample size adequacy, despite significant findings in most parameters. Finally, our results were calculated from only a single simulated game and may not reflect the changes in CoM and joint angles throughout the playing season.
Conclusion
The angles of knee flexion and hip flexion were increased, the CoM position was shifted forward and downward, and the CoM variability increased in all directions after the simulated game with repetitive pitching. Neither pitching accuracy nor velocity decreased after the simulated game. Therefore, decreases in pitching accuracy or pitching velocity might not be used as the sole indicators for player substitution and injury prevention. Baseball pitchers could modify the control of CoM and lower-extremity joint angles to maintain their pitching performance. Compensatory strategies should be monitored to prevent sports injuries in baseball pitchers due to overuse.
Footnotes
Acknowledgements
The authors thank Wen-Chieh Yang, Che-Hsiu Chen, and Chung-Yu Chen, for their support in collecting data. This work was supported by the National Science and Technology Council (NSTC 112-2425-H-028-002).
The authors report no potential conflicts of interest in the development and publication of this article.
This work was supported by the National Science and Technology Council, Ministry of Science and Technology, Taiwan, (NSTC 112-2425-H-028-002).
