1. Introduction
Lower extremity biomechanics are known to influence the kinetic chain in baseball pitching [
1,
2,
3,
4]. Coordinated kinetic chain movements progress up from the feet and lower extremities through the trunk and throwing arm and culminate in the ball release. Efficient distal (lower extremity)–proximal (trunk)–distal (throwing arm) energy transfer depends on how effective momentum is developed and transferred between the segments, beginning with the larger lower body distal segments that achieve peak linear and angular velocities first followed by the lighter distal throwing arm segments that reach peak linear and angular velocities in succession [
5,
6,
7,
8]. Changes in the lower extremity joint power are known to affect the throwing arm kinematics and kinetics, which are reportedly associated with altered trunk–throwing arm momentum sequencing without affecting the ball velocity [
9,
10,
11,
12,
13,
14]. An increased lead leg knee extension concomitant with a longer stride length have reportedly been associated with higher fastball velocities among less adept youth pitchers [
15,
16] whereas proficient pitchers maintain the ball speed despite altered ground reaction force profiles resulting from stride length changes [
17,
18]. The evidence suggests that longer strides have a detrimental physiological cost [
19] yet may be beneficial mechanically in mitigating throwing arm stress [
20,
21]. Conversely, shorter strides reduce timing in the generation phase and prolong the brace-transfer, which evoke inefficient pitching mechanics in preparing the throwing arm for a maximal external shoulder rotation [
22]. Increased throwing arm lag prior to the acceleration phase with shortened strides, concomitant with a greater trunk transverse angular momentum, may increase the strain on the shoulder and elbow [
21] and exacerbate dominant forearm fatigue [
23], which has been associated with a 36 X increase risk for shoulder and elbow injuries [
24].
To explore further, this paper is a continuation of our comprehensive single cohort research study examining whether altering the stride length influences the overhand baseball throwing mechanics [
19] with significant changes in temporal, physiologic, ground reaction, momentum, grip strength, and performance profiles having already been reported [
17,
18,
19,
20,
21,
22,
23]. Whether the stride length adaptations affect the bilateral knee coordination and sagittal link-segment biomechanics without influencing the ball velocity in baseball pitching remains conjecture. Therefore, a secondary analysis of the sagittal plane stride (lead) and drive (trail) leg knee joint dynamics (kinematics and kinetics) in response to altered stride lengths [
18] was used to determine whether the knee joint moments and powers are sensitive to the stride length changes and alter the pitching delivery that may predispose players to an increased risk of lower body and throwing arm injuries.
In the context of this investigation, it was hypothesized that closed kinetic chain lower extremity biomechanics were altered in response to changes in the stride length. Greater drive knee extension dynamics (kinematics, kinetics, and joint powers) were expected with increased strides during the generation phase of pitching (peak knee height to stride foot contact). Likewise, a greater stride leg knee flexion was expected to occur earlier in the pitching cycle (closer in proximity to the stride foot contact) followed by greater stride knee extension dynamics during the acceleration phase (maximal external rotation to the ball release) for shorter strides.
Our findings emphasize that stride length is an important biomechanical parameter that impacts on drive and stride knee dynamics during the pitching delivery, from which the power flow up the link-segment model regulates the angular joint mechanics through to the trunk, shoulder, elbow, wrist, and hand [
2,
25,
26,
27,
28,
29].
2. Materials and Methods
A separate secondary analysis of the existing data from the original single cohort was performed to test the a priori hypothesis that bilateral knee joint kinetic chain biomechanics in response to stride length compensations alter the overhand throwing mechanics. Protocols for the subject recruitment, motion capture, and data post-processing have been previously published [
17,
18,
19,
20,
21,
22,
23].
2.1. Participants
Nineteen healthy and skilled competitive pitchers (aged 18.63 ± 1.67 years, height 1.84 ± 0.054 m, mass 82.14 ± 0.054 kg) from collegiate and high school seasonal travel programs from across Western New York participated. Each gave informed consent or parental consent was granted for minors in accordance with the Institutional Review Board.
2.2. Experimental Design
A randomized cross-over design required pitchers to throw two 80 pitch simulated games on separate days beginning with either: (1) a 25% over-stride (OS); or (2) a 25% under-stride (US) from the desired stride length (DSL). The players were crossed over to the alternate condition after a minimum of 72 h rest had elapsed, which is the safety protocol of Major League Baseball (MLB) for 60 or more accumulated pitches per day.
2.3. Data Collection
Height and weight were measured, then retroreflective markers were affixed bilaterally and reflective tape was secured to the baseball to track whole-body motion and the instant of the ball release. Three-dimensional marker trajectories were reconstructed using an 8 camera Vicon MX20 system (Vicon Motion Systems, Centennial, CO, USA) integrated with two floor-mounted force platforms (KISTLER Corp., Amherst, NY, USA) aligned in a series so that the drive and stride legs contacted the opposing force platforms. The kinematic and force data were recorded synchronously, sampling at 240 Hz and 960 Hz, respectively.
2.4. Stride Length Determination Preceding and during the Simulated Game Conditions
Motion capture commenced with a progressive throwing warm-up, with the last 5 warm-up pitches thrown at 100% effort. The two highest velocities were averaged and the respective desired stride lengths were determined from the kinematic data. Overhand throwing was standardized using the stretch delivery.
The stride leg peak knee height was identified as the highest vertical displacement during the wind-up. The drive and stride leg ground reaction forces were normalized to the body weight, from which the stride foot contact was determined when the vertical ground reaction force exceeded a 5% body weight. The desired stride length was defined as the mean horizontal distance between the drive foot calcaneus at a peak knee height to the stride foot calcaneus at the stride foot contact. Thereafter, the desired stride length was adjusted ±25% to challenge the throwing mechanics. The areas over the force platforms were marked to indicate the drive foot and stride foot placement for both under- and over-stride conditions, where participants were encouraged to contact the targets during the simulated games. The 24% difference in stride between the conditions (normalized as a percentage of body height) was determined a priori and fell between 50–80% of the body height (as measured during the warm-up), which was representative of collegiate and professional pitchers [
30,
31,
32]. All stride lengths were expressed in meters and the ±25% stride conditions differed statistically from the desired stride [
17].
After acclimatizing to the adjusted stride, 20 overhand throws at 100% effort were completed per inning with a ratio of 3 fastballs to 1 change-up for both game simulations. A time of 20 s between pitches (which adhered to MLB regulations), 9 min between innings, and 5 warm-up throws were allocated before each inning. The testing ceased after the 80th delivery, with approximately 130–140 throws accumulated per testing condition. The throwing hand and baseball velocity were tracked at the ball release with a professional radar gun (Jugs Sports, Tualatin, OR, USA) to encourage fastballs to be thrown maximally. Over the 80 throws, the two highest velocities identified from the radar data during the first and last innings (4 trials) were selected for the analysis to account for the influence of physical exertion from pitch accumulation, from which the means were derived.
2.5. Data Management and Post-Processing of the Kinematic and Kinetic Data
Visual 3D (C-Motion Inc., Rockville, MD, USA) was used for the kinematic and kinetic data analysis. The marker trajectories and ground reaction force data were filtered using a second-order bidirectional Butterworth low-pass filter at 13.4 Hz [
1,
2,
32] and 40 Hz, respectively. The overhand throwing cycle from the peak knee height to the ball release was time-normalized to 100%. Hallmark events in the pitching cycle were identified: (i) peak knee height; (ii) stride foot contact; (iii) maximal external shoulder rotation; and (iv) ball release. The ball release was coincident with the peak linear hand velocity, which was verified by visually inspecting the kinematic ball and hand trajectories. Therefore, the peak linear hand velocity was used to terminate the throwing cycle.
Three phases were defined between the hallmark events: (i) generation: from peak knee height to stride foot contact; (ii) brace-transfer: between the stride foot contact and maximal external rotation; and (iii) acceleration: within the maximal external rotation to the ball release. Lower extremity kinematic and kinetic profiles were derived using Euler equations of motion [
3,
18,
19]. Inverse dynamics were used to derive the net internal joint moments and powers over the time-normalized pitching cycle by integrating normalized GRF data synchronized with kinematic data; both the propulsive and bracing GRF profiles have been reported elsewhere [
8]. Sagittal knee joint kinematics and kinetics were obtained at seven instances: peak knee height, generation phase, stride foot contact, brace-transfer phase, maximal external rotation, acceleration phase, and ball release.
Despite a ±25% stride adjustment, the ball velocity remained unchanged [
17] over the accrued innings (mean fastball velocity 123.5 ± 7.98 km/h vs. 122.7 ± 7.19 km/h and mean change-up velocity 109.1 ± 10.1 km/h vs. 106 ± 9.09 km/h).
2.6. Statistical Analysis
Statistical analyses were performed using SPSS 19 (SPSS Inc., Chicago, IL, USA) utilizing paired t-tests for planned pairwise comparisons between over- and under-stride single or double support lower body dynamics at hallmark events and phases. The statistical significance, determined a priori, was set at p ≤ 0.05 for all statistical tests. Effect sizes (ES) were derived to describe the standardized mean differences using Cohen’s d convention as a benchmark to highlight the magnitude of the effect (d effects: small ≥ 0.20, medium ≥ 0.50, large ≥ 0.80).
3. Results
The drive knee dynamic data presented in
Table 1 and
Figure 1A depict that the knee flexion was significantly greater during the generation phase (
p ≤ 0.001, ES = 1.13) followed by a greater drive knee extension during brace-transfer (
p ≤ 0.001, ES = 1.11) with longer strides. Increased stride lengths exhibited a greater drive knee extensor velocity at stride foot contact (
p ≤ 0.05, ES = 0.82), a lower mean flexor velocity during the generation phase (
p ≤ 0.05, ES = 0.36), and greater drive knee extensor moments during brace-transfer (
p ≤ 0.001, ES = 2.83). In contrast, a greater drive knee flexor velocity was observed with shorter strides during the acceleration phase (
p ≤ 0.05, ES = 2.37). Drive knee extensor moments were observed for increased strides whereas shorter strides demonstrated drive knee flexor moments at MER (
p ≤ 0.001, ES = 0.35). Throughout the acceleration phase, knee moments between the stride length conditions demonstrated differences in power generation and absorption. Pitching with longer strides evoked an eccentric knee extensor action that absorbed power by slowing the rate of drive knee flexion whereas under-stride-generated power was evidenced by concentric knee flexor moments that increased the rate of drive knee flexion (
p ≤ 0.001, ES = 2.56).
A greater stride knee extension was observed during brace-transfer (
p ≤ 0.001, ES = 4.17) and the acceleration phase (
p ≤ 0.05, ES =1.80) when pitching with longer strides. However, the angular velocities were no different at any point in the pitching cycle between the stride conditions (
Table 2). Overall, shorter strides resulted in greater stride knee extensor moments at a maximal external rotation and acceleration phase with a moderate to large effect size, respectively (
p ≤ 0.05), with a greater mean joint power generated during the acceleration phase (
p ≤ 0.001, ES = 3.26) (
Table 2 and
Figure 1B).
4. Discussion
The baseball pitch is one of the most dynamic of athletic skills. However, the complex interaction between stride length modification as a compensatory biomechanical response to central and peripheral fatigue on pitching mechanics may amplify shoulder and elbow joint torques and exacerbate throwing arm injuries. Over the course of a game, adjusting the pitching mechanics may be a reactive response to a cumulative workload (number of pitches and aggregate innings), as evidenced by altering the stride length that elicits compensatory lower extremity mechanics followed by the trunk and throwing arm in succession. Biomechanical changes to maintain or increase fastball velocities can affect forward propulsive and braking ground reaction force parameters, which may be a better indicator of exertion that can be monitored [
19]. Given the paucity of research examining coordinated drive and stride leg intersegmental knee joint dynamics (kinematic and kinetic parameters) in baseball pitching, our retrospective analysis of the existing data from an original single cohort during simulated flat ground baseball game pitching confirmed our hypothesis. The net internal knee joint moments and powers appeared to be sensitive to the two stride strategies along with different drive and stride knee angular dynamics, evident in the respective generation, brace-transfer, and acceleration phases (
Table 1 and
Table 2). Our analysis focused solely on closed kinetic chain biomechanics, with the drive leg evaluated over the span of the entire pitching cycle whereas the stride leg (brace knee dynamics) commenced from the stride foot ground contact through to the ball release for both deliveries.
Throughout the pitching cycle, both the drive and stride legs regulate the respective propulsion and bracing ground reaction forces [
18]. Manipulating the stride length (25% under- and over-stride) led to a statistically different stride foot contact as well as duration of single and double supports during the generation and brace-transfer phases between the conditions [
22]. This was concomitant with altered propulsive and bracing ground reaction forces [
18], thereby affecting the drive and stride knee joint kinematics and kinetics. Longer strides reportedly prolonged the propulsive generation phase prior to stride foot contact (80% of the pitching cycle) whereas shorter strides saw the generation phase decrease to 73% of the pitching cycle [
22], consequently altering the knee dynamics. Accordingly, a greater drive knee extensor effort in the propulsion (generation phase) with longer strides was substantiated, concomitant with greater stride knee extensor moments in bracing with reduced strides. The shorter strides delayed the drive leg knee extension mechanics, with associated increases in the stride knee joint powers in bracing during the acceleration phase of throwing.
The longer strides evoked greater drive knee extension and joint power throughout the pitching cycle (maximizing propulsion), with a concomitant increase in the stride knee extension for effective bracing (
Figure 1A). In contrast, shorter strides required less drive knee extensor effort, as evidenced by lower extensor knee moments during the generation phase to propel the body (
Figure 1A). In addition, the longer strides positioned the drive knee further into flexion during the generation phase, which appeared to enhance the extension velocity moving into the stride foot contact, illustrating an improved stretch-shortening of the quadriceps (
Figure 1A). Greater drive knee excursions from flexion to extension with an increased stride length throughout the generation phase is thought to augment extensor preload, drive knee extension, and angular velocities as well as moments in maximizing propulsion prior to the onset of trunk rotation. Pitching with longer strides is thought to mediate a greater efficiency in trail leg braking; this was evidenced by greater drive knee extension and extensor moments during the brace-transfer phase and lower overall stride knee extensor kinetics (
Figure 1B). The opposite was observed with shorter strides as the drive knee extensor moments were lower during the brace-transfer phase (reduced propulsion). The implication of changing the stride length on the trailing leg mechanics in overhead pitching likely alters the inertial properties and momentum sequencing, culminating in trunk–throwing arm momentum transfers. This was exemplified by greater hind knee flexion and flexor moments observed during the brace-transfer phase with shorter strides, which perhaps decreased the mass moment of inertia needed to speed up the proximal rotation. Reduced stride lengths reportedly disrupt trunk–throwing arm momentum sequencing, indicative of an increased trunk transverse momentum that potentially exacerbates throwing arm lag, which mediates greater throwing arm kinetics [
21].
The lower drive knee extensor dynamics observed with shorter strides concomitant with reduced hind leg inertial properties late in the brace-transfer phase necessitated greater stride knee extensor moments during acceleration in bracing. This suggested a greater physical effort in stride leg bracing for shortened strides throughout acceleration, with mean stride knee extensor moments nearly 83% of the body weight x height in braking and approximately 13% higher compared with pitching with longer strides. The greater maximum stride knee extensor moments and velocities observed during bracing with shorter strides (
Figure 1B) were indicative of a greater knee extensor power generation, suggesting that increased stride knee bracing may be compensatory owing to reduced drive knee propulsion during the generation phase. As a result of atypical bracing, throwing arm injury risks can be exacerbated by altered kinetic chain interactions through stride length compensations.
As this study was part of a comprehensive single cohort full-body biomechanical analysis building on our previous research, it may—when combined with recent evidence—provide additional evidence-based criteria that can be used to optimize the stride length, evaluate lower body fatigue, and prevent upper extremity injuries among baseball pitchers. Each study is considered to be independent because they differ in their research question, hypothesis, analytic methods, and conclusions; however, the findings are complementary in supporting how coordinated 3D lower extremity biomechanics are altered in response to changes in stride length, influencing the kinetic chain that potentially induces compensatory throwing mechanics and predisposes pitchers to a greater risk of injury.
The sequential order of the post-hoc secondary analysis suggested that longer strides have a detrimental physiological cost, manifested by throwing arm fatigue [
23] concomitant with significantly higher perceived exertion, heart rate, and metabolic responses coincident with muscle inflammation and pain [
19,
33] yet may be beneficial mechanically in mitigating throwing arm stress. Despite ball velocity performance remaining unchanged [
17], significantly different propulsive and bracing ground reaction force profiles in the magnitude and timing of peaks [
18], the linear and angular momentum transfer [
20,
21], temporal profiles [
22], and grip strength [
23] have subsequently been reported, augmenting the initial findings [
19]. Shorter strides reduced the timing in the generation phase and prolonged brace-transfer, which evoked inefficient pitching mechanics in preparing the throwing arm for a maximal external shoulder rotation [
22]. Increased throwing arm lag prior to the acceleration phase with shortened strides, concomitant with a greater trunk transverse angular momentum, increased the strain on the shoulder and elbow [
21] and exacerbated dominant forearm fatigue [
23], which has been associated with a 36 X increase in risk for shoulder and elbow injuries [
24]. As the ball velocity was maintained between the stride conditions [
17,
18], the pitchers appeared to be capable of generating a variable power between the drive and stride knee extensors without adversely affecting the performance.
The knowledge gained offers new insights into compensatory stride length adaptation evoked by cumulative physical exertion (amassed throwing workload over time or by coaches facilitating adjustments) on respective drive and stride leg dynamics throughout the pitching cycle. Perhaps a concomitant assessment of in-game ball velocity and pitch count metrics as well as simultaneous biomechanical analyses of stride length may be more effective in monitoring physical exertion (accumulated pitching workload) as underlying compensatory lower body biomechanics to maintain peak ball velocities go unnoticed and can be detrimental to throwing arm health.
Limitations
Pitching in laboratory settings does not replicate an actual competition; therefore, interpreting the findings must be scrutinized pragmatically. Notably, the indoor lab setting imposed non-regulation dimensional constraints and despite the absence of a mound where participants threw without wearing cleats, flat ground pitching is commonly practiced and reinforces the importance of this study. However, a comparison of mound and flat ground collegiate throwing profiles demonstrated similar biomechanical patterns, evidenced by an equivalent shoulder internal rotational torque and medial elbow load when throwing at maximum effort [
34,
35]. However, adolescent lower extremity kinematic movement profiles differ [
36], with differences evident in the timing of events in the normalized pitching cycle relative to the lead foot contact as well as statistically higher throwing arm kinetics when throwing from the mound in comparison with flat ground conditions.
In this study, the pitching sequence was time-normalized from the peak knee height (0%) to the ball release (100%) to include the generation phase; otherwise, the delivery mechanics commencing at the stride foot contact would only have captured the brace-transfer and acceleration phases. However, comparisons across studies may be difficult because the kinematic and kinetic time series data as a function of the normalized pitching cycle are typically expressed as a percentage of the delivery (0–100%), commencing with the stride foot contact (0%) and culminating in the instant of the ball release (100%) [
37].
A simple randomized cross-over design (AB/BA) was utilized to control for order effects. In randomizing the allocation to the sequence of the two stride conditions given consecutively, the advantage was that each subject served as their own control allowing for contrasting responses to condition A with B. Removing patient variations in this way makes cross-over trials potentially more efficient and a smaller number of participants are required. The disadvantage of this methodological design was the inability for comparing the desired stride length data with the ±25% stride conditions. To enable a comparison with the desired stride length, more complicated counterbalanced cross-over designs are necessary. The complexity of having 3 stride conditions (under-stride, desired stride, and over-stride) necessitates 3! = 6 possible sequences, requiring a larger number of subjects for each possible group. This counterbalanced design precludes 6 independent groups crossed over to respective stride conditions, as shown in
Table 3.
This may serve as the basis for future investigators utilizing more complex research design methods to substantiate these kinetic changes with desired stride length profiles.