Throwing is critical for success in many sports, yet it carries an inherent risk of injury (10,18,23). This risk increases with the volume of throws performed per game and per year (15). Fatigue that occurs during a bout of repetitive throwing is thought to exacerbate this injury risk, with an increased prevalence of shoulder pain evident in pitchers who continue to throw in the presence of “arm fatigue” (10).
Consequently, the ability of coaching staff to determine when an athlete is fatigued during the course of a bout of repeated throwing is critical for injury prevention within this population. The early detection of arm fatigue would allow for enhanced clinical decision making to take place, providing an evidence base for an individual's pitching volumes and restrictions, as well as informing decisions as to when an athlete should be removed from the game.
As a result, much research has sought to identify specific parameters that may be indicative of arm fatigue. Throwing velocity (4), arm soreness (16,25), shoulder muscle strength (11,25), throwing arm proprioception (20,21), muscle damage (16), and throwing kinematics (12) have been shown to be affected by a single bout of repeated throwing. Problematically, however, these studies did not use a randomized control design, raising doubts as to whether throwing was the actual cause of the observed changes, or merely a result of taking repeated measures over time. To accurately gauge the level of arm fatigue experienced by a throwing athlete, support staff should closely monitor only those parameters that are negatively affected by repeated throwing itself. Currently, it remains unclear as to which variables are indicative of arm fatigue and should be monitored during bouts of throwing. Consequently, the aim of this study was to determine key markers that are indicative of arm fatigue, to guide monitoring practices of staff working with throwing athletes who could benefit from the early detection of arm fatigue.
We conducted the study with the following hypothesis: That throwing velocity and accuracy as well as shoulder proprioception and soreness in the throwing arm would be significantly and negatively affected by throwing-specific exercise and therefore be presented as key indicators of arm fatigue that should be monitored closely during bouts of repeated throwing.
Experimental Approach to the Problem
To determine the key indicators of arm fatigue, this study used a randomized crossover design comparing the effects of throwing-specific and running-based exercise on potential markers previously identified in the literature. The assumption was that factors, which are negatively affected by throwing exercise specifically (rather than general exercise), would be the most sensitive markers of arm fatigue and therefore ought to be the variables monitored by coaches of throwing athletes.
Thirteen junior elite baseball players participated in the study (19.6 ± 2.6 years, 17–23 years, 1.80 ± 0.05 m, and 75.5 ± 7.4 kg), subject to the condition that they had played baseball at the state level within the past 18 months and were currently selected in a state-level high-performance baseball program. Appropriate consent was obtained pursuant to law because all participants were informed of the experimental procedures involved before providing written consent to participate in the study. Parental or guardian consent was also obtained in the event the participant was younger than 18 years of age. They also completed a questionnaire regarding injury and playing history. All research was in compliance with the ethical guidelines for human research laid down by the Australian NHMRC, and institutional board approval was provided by the University of Sydney Human Research Ethics Committee.
Testing took place during the off-season phase of competition. The participants were required to attend the facility on 2 occasions separated by a minimum of 7 days. Testing was conducted within an indoor laboratory, and each session took place at the same time of the day. The participants were instructed to maintain a high level of hydration and to refrain from eating within 2 hours of the testing session.
Shoulder Proprioception Assessment
On arriving at the facility, the participants were assessed bilaterally for active shoulder proprioceptive acuity using a method previously developed (24). Briefly, we used a specifically designed Active Movement Extent Discriminating Apparatus (AMEDA) (13), consisting of a computer-controlled motor mounted on an adjustable hoist, driving a geared shaft with a circular wooden disc at its end (RS Components Pty. Ltd., Silverwater, Australia). Standing in 90° of shoulder abduction, elbow flexion, and shoulder external rotation, the participant's arm was rested on a padded rest. This position was defined as position 0. The motor could move the disc anteriorly from this position to 1 of 5 different locations. Additional random movements of the shaft were made during repositioning to prevent any use of auditory cues as to the end location of the disc. Each position was approximately 1.5 cm apart and elicited 74° of external rotation at position 5 and 85° at position 1. The left and right arms were tested and the order was randomized. After the nature of the test was explained to the participant, 21 practice trials were performed with 3 trials to each of the 5 locations and a further 3 trials to the 2 terminal locations. The participants then performed a randomized sequence of 50 trials under test conditions comprising of 10 trials for each of the 5 positions. During each practice and test trial, the participants were instructed to face forwards, with the elbow in full extension. When the disc was in its final position, the participants were required to flex the elbow to 90° then externally rotate the shoulder until the hand contacted the disc. The participants estimated the position of the disc (from 1 to 5) immediately after hand contact with the disc. The examiner, seated behind the participant to ensure correct movement, entered each response without the knowledge of the actual position. Testing took approximately 30 minutes. The intraclass correlation (ICC) and the SEM for test-retest reliability for the proprioception measure were 0.86 (95% CI, 0.52–0.96) and 0.03 receiver operator characteristic (ROC) units, respectively.
Throwing Velocity and Accuracy Assessment
The participants then performed a general warm-up routine consisting of 5–10 minutes of moderate intensity running followed by 5–10 minutes of general stretching of the major muscle groups. This was followed by 10–15 minutes of throwing, which began at light intensity before progressing to high intensity at completion (5,6).
After the warm-up procedure, the participants performed a total of 10 throws from the set position (1 forward stride permitted before release) toward a target located 20 m from the participant's back foot. The target consisted of a white circle measuring 7.0 cm in diameter, painted onto a black rubber mat that was suspended on a specifically designed metal frame (Figure 1). The center of the target circle was at a height of 0.70 m from the ground. Using a regulation size (approximately 6.7 cm diameter) and weight (142 g) baseball, 10 throws were performed at 100% of maximal throwing velocity (MTV), measured with a Cordless Speed Radar Gun (Jugs Corporation, Tualatin, OR, USA) positioned behind the participant. Ten seconds separated each throw to minimize the fatigue effect. A video camera (SONY Corporation, Tokyo, Japan) recording at 50 Hz was used to determine the point of ball contact relative to the target. The ICC and the SEM for test-retest reliability for the MTV test were 0.98 (95% CI, 0.93–0.98) and 0.4 m·s−1, respectively, and for the accuracy test were 0.87 (95% CI, 0.54–0.97) and 6.7 cm, respectively.
After the throwing accuracy test, the participants were randomly allocated to perform either a throwing-specific (THROW) or running exercise (RUN) protocol. A crossover design was used such that the participants completed the alternate exercise protocol during the second testing session. THROW consisted of 60 near maximal effort throws (>95% MTV as determined by a cordless radar gun) over a distance of 20 m toward no specific target, each throw being separated by 10 seconds. The total time for this protocol was 10 minutes. RUN consisted of one 20-m shuttle run test (beep test) to exhaustion (9). The total time for this protocol was dependent on the fitness level of the individual and ranged between 9 and 12 minutes.
Immediately before and after exercise, the participants were required to subjectively rate from 0 to 10 their level of soreness located specifically in the throwing arm using the universal pain assessment tool (1). They were also asked to subjectively rate their level of perceived exertion using the Borg 6–20 scale (2).
After the exercise protocol, the participants rested for a period of 3 minutes and were then reassessed for throwing accuracy and shoulder proprioception acuity as described above.
Footage of each throw was filmed and imported into a measurement analysis software package (Swinger Pro; Webbsoft, Scoresby, Australia). A still frame of the time of contact between the ball and target was generated. The distance between the center of the ball and the target center was then calculated by measuring the number of pixels between the ball and target before converting this to a length in meters using the calibration frame (Figure 1).
Throwing accuracy was determined by assessment of specific parameters defined previously (7) including total error (E), absolute constant error (ACE), variable error (VE), vertical error (Y), and horizontal error (X). The E was calculated as the average distance between each throw and the target. The ACE, a measure of bias, was calculated as the distance between the typical throw (centroid) and the target, with the typical throw defined as the centroid calculated by averaging the X and Y values of each throw in a set. Variable error, a measure of consistency, was calculated as the average distance between each throw and the typical throw. The X and the Y errors were calculated as the average distance between each throw and the target in the X and Y directions, respectively.
For each subject, ROC graphs were generated for the shoulder proprioception data obtained for the dominant arm. The area under the curve (AUC) gives an accuracy of discrimination measure (19). An AUC of 1.0 represents perfect accuracy, whereas a score of 0.5 represents chance performance. The AUC was calculated from positions 1 and 5 of the AMEDA.
The hypothesis was tested by comparing the effect caused by throwing-specific exercise to that caused by running-based exercise on each of the factors of interest. To do this, the data set was subjected to ANCOVA (22), where the dependent variable was the change score, calculated as the postscore minus the prescore. The independent variable was the type of exercise (THROW vs. RUN), and the covariate was the score at baseline. The assumptions for ANCOVA were met. Outcomes were grouped into velocity (n = 3), accuracy (n = 5), subjective measures (n = 2), and proprioception (n = 2) variables, with the Holm-Bonferroni adjustment applied. T-tests were also conducted on each dependent variable to determine time effects for each of the exercise conditions. Pearson's correlation coefficients were calculated on baseline and change scores to determine the relationship between shoulder proprioceptive acuity, MTV, and throwing accuracy. Statistical significance was set at p ≤ 0.05.
Mean and peak MTV were significantly reduced after THROW and RUN protocols to a similar degree (p ≤ 0.05). Velocity variability was not significantly different after either exercise bout (Table 1).
Total error and X were significantly increased after THROW (p ≤ 0.05), but unchanged after RUN, leading to a significant exercise effect, (p ≤ 0.05). Absolute constant error was significantly increased after THROW (p ≤ 0.05) and significantly reduced after RUN (p ≤ 0.05) leading to a significant exercise effect (p ≤ 0.05). Variable and Y errors were not affected by the THROW or RUN protocol (Table 1).
Rating of perceived exertion and throwing arm soreness increased significantly after RUN and THROW (p ≤ 0.05). The increase in RPE was significantly greater after RUN compared with THROW (p ≤ 0.05), whereas the increase in arm soreness was significantly greater after THROW compared with RUN (p ≤ 0.05). Proprioception was not significantly affected by RUN or THROW. No significant correlations were found between proprioception and throwing velocity or accuracy (Table 2).
This is the first study to compare the effects of throwing-specific and general exercise on potential markers of arm fatigue in throwing athletes. As hypothesized, the results confirmed previous findings (4,17,25) and demonstrated that throwing velocity, throwing accuracy, and throwing arm soreness were negatively affected by throwing exercise. Interestingly, throwing velocity was reduced to a similar degree after both throwing and general exercise bouts, whereas accuracy and arm soreness were affected to a far greater degree by throwing exercise. Although throwing velocity may prove to be a useful indicator of general fatigue, these results do not suggest that it is a sensitive marker of arm fatigue because the true cause of the observed reduction in velocity cannot be confidently attributed to the throwing activity itself. Alternatively, changes in throwing accuracy and arm soreness are attributable to throwing-specific exercise because accuracy was only reduced after throwing exercise, and the increase in arm soreness was significantly greater after THROW than after RUN. As a result, these 2 measures present themselves as key indicators of arm fatigue with the potential to act as early detectors of throwing-induced injury risk.
Although no studies have described accuracy changes after throwing-specific exercise, the results of this study confirm anecdotal evidence from the baseball coaching community which views accuracy as a sensitive predictor of arm fatigue. Interestingly, however, within this commonly held belief, vertical accuracy is thought to be the variable of interest, with throws that miss high in the strike zone suggested to be more indicative of fatigue. In this study, horizontal rather than vertical accuracy worsened over the course of the throwing bout, with the participants less able to hit the target in the left to right plane. The specific cause of this finding remains unclear. Previous investigations have shown repeated throwing to result in significant changes to a number of kinematic and kinetic parameters such as lateral trunk tilt (4), maximal external shoulder rotation, knee angle at release, maximal shoulder and elbow distraction force as well as horizontal shoulder adduction torque (12). Although repeated throwing potentially reduces throwing accuracy by creating deviations away from typical kinematic and kinetic sequences (8), such a conclusion is beyond the scope of this study and provides stimulus for future investigation.
Interestingly, and in contrast with previous research, this study does not support the monitoring of shoulder proprioception as an indicator of arm fatigue. Although previous investigations have shown reduced shoulder proprioception after throwing-specific exercise (20,21), this study found no such relationship. We attribute this to differences in the complexity of the tasks used. This study involved 3-dimensional unconstrained throwing, a highly complex and coordinated skill, whereas previous investigations into proprioception used constrained throwing in which the lower limbs were physically restrained or minimally involved in the throwing motion. Although the role of shoulder proprioception in throwing performance and injury risk cannot be ruled out, there is little evidence to suggest that the shoulder proprioception of throwing athletes should be routinely monitored at this time. This position is reinforced by the complex and time-consuming nature of this type of assessment. Given the poor feasibility of proprioception measurement, combined with its lack of sensitivity to detect changes in arm fatigue, this monitoring practice ought not to be adopted at this time.
In addition to monitoring practices, the current results have significant implications for the training of throwing athletes. The demands of throwing-specific exercise were demonstrated to place unique stresses on the body, significantly increasing arm soreness and reducing performance in terms of throwing velocity and accuracy. These data highlight the need for throwing athletes to engage in throwing-specific conditioning and prehabilitative exercises (3,6,14) to build their resilience to throwing-induced arm fatigue and attenuate the performance decrements that result from this type of activity.
Overall, this study identified a number of key variables that are indicative of fatigue during a single bout of repeated throwing. Throwing velocity, throwing accuracy, and the degree of soreness in the throwing arm may provide insight into the level of overall fatigue and, therefore, injury risk within this population. More specifically, however, throwing accuracy and arm soreness are presented as indicators of arm fatigue, whereas velocity may be more indicative of general fatigue. Consequently, the routine monitoring of these particular variables is warranted and may provide a greater evidence base upon which clinical decisions can be made. Coaches should routinely ask their pitchers to rate their level of arm soreness using a scale similar to that used in this study. When available, coaches should measure the ability of their pitchers to throw accurately using ball-tracking systems. In lieu of this technology, coaches should observe the accuracy of the pitcher, as well as obtain subjective ratings of accuracy from the catcher and pitcher between innings. Information regarding throwing velocity, throwing accuracy, and arm soreness should be used to make decisions about what pitching restrictions an individual should be placed on and when they should be removed from the game. Greater weighting should be placed on arm soreness and throwing accuracy during this decision-making process, as velocity may be more indicative of general rather than arm fatigue. Interestingly, this study does not support the practice of shoulder proprioception monitoring as a marker of injury risk at this time.
The results of this study do not constitute endorsement of the product by the authors of the National Strength and Conditioning Association. No external funding was received for this study. There are no conflicts of interest to declare.
1. Baker C, Wong D. Q.U.E.S.T.: A process of pain assessment in children. Orthop Nurs 6: 11–21, 1987.
2. Borg G. Perceived exertion as an indicator of somatic stress. Scand J Rehabil Med 2: 92–98, 1970.
3. Escamilla R, Fleisig G, Yamashiro K, Mikla T, Dunning R, Paulos L, Andrews J. Effects of a 4-week youth baseball conditioning program on throwing velocity. J Strength Cond Res 24: 3247–3254, 2010.
4. Escamilla RF, Barrentine SW, Fleisig GS, Zheng N, Takada Y, Kingsley DS, Andrews JR. Pitching
biomechanics as a pitcher approaches muscular fatigue during a simulated baseball game. Am J Sports Med 35: 23–33, 2007.
5. Freeston J, Ferdinands R, Rooney K. Throwing velocity and accuracy in elite and sub-elite cricket players: A descriptive study. Eur J Sport Sci 7: 231–237, 2007.
6. Freeston J, Rooney K. Progressive velocity throwing training increases velocity without detriment to accuracy in sub-elite cricket players: A randomized controlled trial. Eur J Sport Sci 8: 373–378, 2008.
7. Hancock G, Butler M, Fischman MG. On the Problem of Two-Dimensional Error Scores: Measures and Analyses of Accuracy, Bias and Consistency. J Mot Behav 27: 241–250, 1995.
8. Hirashima M, Kadota H, Sakurai S, Kudo K, Ohtsuki T. Sequential muscle activity and its functional role in the upper extremity and trunk during overarm throwing. J Sports Sci 20: 301–310, 2002.
9. Leger L, Mercier D, Gadoury C, Lambert J. The multistage 20 metre shuttle run test for aerobic fitness. J Sports Sci 6: 93–101, 1988.
10. Lyman S, Fleisig GS, Waterbor JW, Funkhouser E, Pulley L, Andrews J, Osinski D, Roseman J. Longitudinal study of elbow and shoulder pain in youth baseball pitchers. Med Sci Sports Exerc 33: 1803–1810, 2001.
11. Mullaney MJ, McHugh MP, Donofrio DM, Nicholas SJ. Upper and lower extremity muscle fatigue after a baseball pitching
performance. Am J Sports Med 33: 108–113, 2005.
12. Murray TA, Cook TD, Werner SL, Schlegel TF, Hawkins RJ. The effects of extended play on professional baseball pitchers. Am J Sports Med 29: 137–142, 2001.
13. Naughton J, Adams R, Maher C. Discriminating overhead
points of contact after arm raising. Percept Mot Skills 95: 1187–1195, 2002.
14. Niederbracht Y, Shim AL, Sloniger MA, Paternostro-Bayles M, Short TH. Effects of a shoulder injury prevention strength training program on eccentric external rotator muscle strength and glenohumeral joint imbalance in female overhead
activity athletes. J Strength Cond Res 22: 140–145, 2006.
15. Olsen SJ. Risk factors for shoulder and elbow injuries in adolescent baseball pitchers. Am J Sports Med 34: 905–912, 2006.
16. Potteiger J, Blessing D, Wilson D. Effects of varying recovery periods on muscle enzymes, soreness, and performance in baseball pitchers. J Athl Train 27: 27–31, 1992.
17. Potteiger JA, Blessing DL, Wilson GD. The physiological responses to a single game of baseball pitching
. J App Sport Sci Res 6: 11–18, 1992.
18. Saw R, Dennis R, Bentley D, Farhart P. Throwing workload and injury risk in elite cricketers. Br J Sports Med 45: 805–808, 2009.
19. Swets J. Measuring the accuracy of diagnostic systems. Science 240: 1285–1293, 1988.
20. Tripp B, Yochem E, Uhl T. Functional fatigue and upper extremity sensorimotor system acuity in baseball athletes. J Athl Train 42: 90–98, 2007.
21. Tripp BL, Boswell L, Gansneder BM, Shultz SJ. Functional Fatigue Decreases 3-Dimensional Multijoint Position Reproduction Acuity in the Overhead
-Throwing Athlete. J Athl Train 39: 316–320, 2004.
22. Vickers A, Altman D. Analysing controlled trials with baseline and follow up measurements. BMJ 323: 1123–1124, 2001.
23. Wheeler K, Kefford T, Mosler A, Lebedew A, Lyons K. The volume of goal shooting during training can predict shoulder soreness in elite female water polo players. J Sci Med Sport 16: 255–258, 2013.
24. Whiteley RJ, Adams RD, Nicholson LL, Ginn KA. Shoulder proprioception is associated with humeral torsion in adolescent baseball players. Phys Ther Sport 9: 177–184, 2008.
25. Yanagisawa O, Miyanaga Y, Shiraki H, Shimojo H, Mukai N, Itai Y. The effects of various therapeutic measures on shoulder strength and muscle soreness after baseball pitching
. J Sports Med Phys Fit 43: 189–201, 2003.