Adolescent tennis players are known to be at a greater risk of low back pain (LBP) and structural injury than their peers (5,24). For example, a 2-yr survey revealed that 38% of a group of 12- to 18-yr-old Swedish players reported a lumbar region injury that resulted in missed training and/or competition (19). Although Australian injury rates have not been published, data from current Tennis Australia athletes suggest a high incidence of low back problems, with 7 of the 41 male players, 12- to 18-yr-old, reporting disabling LBP in 2012. For these players, each LBP incident resulted in an average of 34 d of missed training, representing nearly half of all days lost as a result of injury. In line with other tennis populations, these injuries seem to be mostly associated with a repetitive stress mechanism (19,36,37), with lumbar pars interarticularis stress reactions in adolescent tennis players common (24).
To date, there has been little research regarding potential mechanisms associated with LBP injury and pain in tennis, hampering the development of both evidence-based preventative and injury management strategies (2,29). LBP is known to be a multifactorial condition, with potential risk factors grouped broadly into psychological, social, and biological domains (4). However, the repetitive ballistic trunk movements required in tennis, which have been linked with the high frequency of pars interarticularis stress reactions in other populations (15), underpins the likelihood of a mechanical etiology in this population (2,23). It has been proposed that the repeated rapid rotation of the lumbar spine during tennis ground strokes and the “hyperextension” during serving may be associated with the high rate of radiological abnormalities in tennis players (2). Given the known relationship between excessive mechanical forces and lumbar vertebrae bone injury (9,24), assessing the magnitude of force absorbed through the lumbar spine in tennis is important to gain an insight into back injuries in this sport. While it is not practical to directly measure force through the lumbar vertebrae during tennis stroke execution, it is possible to estimate lumbar region kinetics in vivo (34), and such analyses have proved useful in other sport populations (13). For example, the analysis of lumbar region kinetics during cricket fast bowling facilitated the understanding of the relationship between lumbar region mechanics and LBP in this cohort (13). To date, detailed lumbar region kinetics has not been reported for any tennis stroke.
The tennis serve is critical to match success and is therefore frequently repeated in both training and match play (21,32). It is thought to generate and therefore require absorption of greater force through the lumbar region than the game’s other strokes, with some evidence of serve type influencing the loading of the joints and musculature of the trunk (10,35). Players execute various types of serves but most commonly use the “flat” serve and “kick” serve as their first and second deliveries, respectively (32). The kick serve has a ball impact location further away from the hitting side of the player (32,35) and has been reported to exhibit heightened EMG activity of the trunk musculature (10) compared with the flat serve. Recently, Sheets et al. (35) compared the single-segment “trunk” kinetics between serve types and found that the kick serve generated greater total back force (34). Unfortunately, however, their study did not report kinetics of the lumbar region. Given that the lumbar region is the site of most spinal pain and radiological abnormalities in adolescent tennis players (2,19), a more thorough evaluation of the lumbar region during the flat and kick serves is required. Therefore, the aim of this study was to explore the relationship between LBP and lumbar region kinetics during the flat and kick serves of symptomatic and asymptomatic elite, adolescent players. A secondary aim of this study was to investigate whether differences exist in lumbar kinetics between the flat and kick serves. Lumbar region kinematics, as well as racquet velocity and the position of the ball at impact, was described to facilitate kinetic data interpretation.
Male players were recruited from the current members of Tennis Australia’s National Academy and Australian Institute of Sport tennis programs. A survey before data collection was performed to confirm subject eligibility and pain/no pain grouping, with 21 subjects aged between 13 and 17 yr selected to participate in the study. Players were included in the LBP (pain) group if they had at least two episodes of LBP aggravated by serving that required treatment, resulted in missed training/competition, and was associated with a lower lumbar spine (L4–L5) injury confirmed by appropriate radiological scanning (MRI/CT) and sports physician diagnosis, in a 2.5-yr window around the time of data collection (12–15 months before or after data collection; Table 1).
Participants were excluded if they had musculoskeletal injury or illness preventing maximum skill execution within 3 wk of testing, had been on a modified training load within 3 wk of testing, or at the time of testing indicated LBP during the serve. One player was unable to perform maximal effort serves because of LBP and was therefore excluded. The control group (no pain) had no history of any disabling LBP incident and were of similar anthropometry and age to the pain group (Table 2). As the pain group tended to have greater years playing experience than the no pain group (mean difference = 1.9 yr, 95% confidence interval [CI] = −0.2 to −3.6 yr), years of experience was incorporated as a covariate in statistical analyses.
The data collection was conducted on a full-sized replica tennis court within The Australian Institute of Sports’ Canberra biomechanics laboratory. Relevant institution ethical approval was granted for this investigation (HR 144/2009) and the necessary parental and athlete consent/assent were obtained. The VICON motion analysis system (Oxford Metrics, Inc.), operated at 250 Hz, was used to collect marker trajectory information of all static and dynamic trials.
After arrival to the laboratory, a participant’s height and mass were recorded. An upper and lower limb, trunk, pelvis (26) and lumbar spine (34,38) marker set, which followed International Society of Biomechanics’ recommendations (41), was affixed to specific anatomical landmarks on each participant using double-sided tape. A single-subject calibration trial was then completed. A subject performed a self-directed warm-up, followed by a series of serving trials from the deuce side of the replica tennis court. Trials were only deemed successful if they landed within a preset target area (1-m2 area bordering the “T” of the deuce service box) and were within 5 km·h−1 of their prerecorded maximal serve velocity, as determined by a Stalker radar gun (Applied Concepts, Inc.), during each serve. Serving continued until three successful flat and three successful kick serves were recorded. Players hit 10–16 flat serves and 7–11 kick serves to record the above successful trials.
Marker trajectories were checked for “breaks” or missing information that can occur as a result of marker occlusion using VICON Nexus Software (Oxford Metrics, Inc.). The breaks were infrequent and less than 20 frames in length. Standard procedures were used to interpolate these missing data. After a residual analysis, the trajectories were filtered with a Woltring filter using a mean square error of 3 (39). Ball and racquet impacts were treated, and position data and output were calculated, as previously detailed (32,33). A customized racquet, upper limb and trunk mathematical model was used to calculate lumbar region kinematics and kinetics (7,8,11). The required upper limb, thorax, and lumbar region segment parameters (i.e., length, mass, moment of inertia, and center-of-mass location) for kinetic calculations were obtained from previously published data (12,27,28), with each participant’s racquet segment parameters calculated before data collection using specialized equipment; the “Prince Tuning Centre” (Prince Tennis, Bordentown, NJ). Lumbar kinematics was calculated from the lumbar region movement relative to the pelvis, using a z–x–y Euler angle decomposition (40). Lumbar moment and force data were computed with a “top-down” (i.e., racquet down to the lumbar region) approach (20). All lumbar region kinematics and kinetics were expressed in the lumbar region coordinate system defined from a vertical axis between the fifth and first lumbar vertebrae markers, a sagittal plane axis between markers placed 5 cm on either side of the fourth and fifth lumbar vertebrae junction, and a frontal plane axis defined as the cross product between the vertical and sagittal plane axes (11). The directionality of the frontal and transverse plane results from the single left-handed player was reversed in order for all results to be described using a positive right hand assumption. Therefore, positive values represented lumbar flexion, right lateral flexion, and right rotation (i.e., toward the racquet arm). For the force data, positive values represented anterior, inferior vertical, and right lateral forces.
A custom LabView program (National Instruments, Inc.) used racquet tip and ball position data to identify four temporal events corresponding to established points of practical interest in the serve (32), namely, ball release, first racquet high position, first racquet low point, and ball/racquet impact (Fig. 1). Temporally normalized time-series data from release until impact were averaged for each player and each group to facilitate qualitative analysis of the lumbar region mechanics and the conceptualization of key kinetic events of interest within the serve. Peak force data from two phases were output, namely, drive (from racquet high point to racquet low point) and forward swing (from racquet low point to impact; Fig. 1).
Ball position at ball–racquet impact was calculated relative to the foot and was normalized to height using the above-mentioned LabView program, whereas the lumbar force and moment data were normalized to body mass. The racquet tip markers’ peak resultant velocity and the output for each trial, along with the absolute peak lumbar angle, angular velocity, joint reaction force, and moment in each plane of movement during both drive and forward swing phases were calculated. Mean data across the participant’s three trials were input into statistical software for analysis (SPSS version 19; SPSS, Inc., Chicago, IL). A series of 2 × 2 mixed-model ANOVA were performed to compare between pain/no pain (between subjects) and kick/flat serve (within subjects) for each variable in each phase of the serve. As aforementioned, years of playing experience was incorporated as a covariate in all analyses. The interaction between pain status and serve type was analyzed and found not to be significant for any variable, permitting grouped results (i.e., pain/no pain for both serves) to be presented. The α value was adjusted a priori to ≤0.01 to account for the multiple comparisons, without excessively inflating the possibility of type 2 errors. The sample size of 7 pain and 13 control participants allowed a 0.79 power to detect a 3.5-N difference between groups, using an α value of 0.01.
Racquet velocity and ball position
There was no significant difference in peak racquet velocity between pain groups or serve types (Table 3), although there was a trend for players to hit the flat serve with higher resultant velocities (mean difference = 2.6 m·s−1, 95% CI = −0.4 to 5.4 m·s−1, P = 0.048). There were differences in ball position at impact between serve types, with players typically impacting the ball further from their racquet side, in the frontal plane, during the kick serves (mean difference = 0.1 cm·h−1, 95% CI = 0–0.2 cm·h−1, and P = 0.008). A further trend for the ball to be positioned closer in front of the players’ bodies during the kick serves was also observed (mean difference = 0.08 cm·h−1, 95% CI = 0–0.1 cm·h−1, and P = 0.032).
Qualitative description of lumbar region mechanics
Qualitative inspection of the lumbar region kinematic and kinetic time-series data of both serve types confirmed the appropriateness of the event markers, with opposing lumbar region movements performed in the drive and forward swing phases. Each player’s time-series data showed that, during the drive phase, the lumbar region was extended, while rotated and laterally flexed toward the racquet arm, and that the instances of peak extension and lateral flexion appeared to coincide. Only a small range of rotation occurred throughout the drive phase, first toward the racquet arm, with initial counterrotation beginning before peak extension/right lateral flexion. Figure 2 illustrates this using grouped data. During the transition between the drive phase and the forward swing phase, the lumbar region began to be rapidly “recoiled” (or flexed with simultaneous rotation/lateral flexion away from the racquet arm). These contrary joint rotations were associated with distinct force and moment profiles. Midway through the drive phase, the lumbar region reached peak extension/lateral flexion while simultaneously absorbing peak left lateral flexion and vertical forces. The forces in all three planes seem to be greater during the drive phase than during the forward swing phase. In contrast, forward swing was associated with large moments in all planes.
The comparisons of the peak angular displacement and velocity of the lumbar region between groups (pain/no pain and serve types) revealed no significant differences in the sagittal and frontal planes during either phase. Differences were evident in the transverse plane, with the pain group typically using less right rotation (toward the racquet arm) during the drive phase than the no pain group (mean difference = 3.5°, 95% CI = 1.2°–5.8°, P = 0.004). A trend for more left lumbar rotation (away from the racquet arm) during the forward swing phase was also evident (mean difference = 3.0°, 95% CI = 0.7°–5.3°, P = 0.023). This is consistent with the systematic offset of pain versus no pain group rotation observed in the time-series data (Fig. 2C). Although no statistical differences in peak angular velocities in any plane were evident, it seemed that the pain group used marginally greater right lateral flexion and rotation angular velocity than the no pain group in the drive phase of both types of serve (Figs. 2E and F).
The analysis of peak forces and moments between the pain and no pain groups (Table 4) indicated that the players with LBP used significantly greater left lateral force during the drive phase (mean difference = 1.5 N·kg−1, 95% CI = 0.5–2.6 N·kg−1, P = 0.01). Lumbar kinetics during the forward swing was generally comparable between the pain and no pain groups. The comparison between serves revealed that the flat serve was associated with significantly greater flexion moments than the kick serve during the forward swing phase (mean difference = 2.7 N·kg−1, 95% CI = 1.4–3.9 N·kg−1, and P < 0.001).
This is the first study to quantify the lumbar region kinetics and kinematics during the tennis serve. Differences were found in the peak lateral forces and lumbar rotation during the drive phase of serves between elite, adolescent tennis players with and without LBP. Differences between the kick and flat serves were also detected in the lateral ball position and peak flexion moment during the forward swing phase.
Pain and no pain comparison
The kinetic data revealed that the main difference between the pain and no pain groups was that the pain group used significantly greater lateral flexion force during the drive phase of both kick and flat serves than the no pain group. The pain group reported peak lateral flexion forces approximately four times body weight, around 50% greater than the no pain group. When compared with other sports like running, the drive phase of the serve is characterized by similar vertical forces but with much higher—in the order of eight times—lateral flexion forces (34). Given that tissue damage results from the absorption of forces beyond tissue tolerance (1), these large lateral flexion forces in the pain group have the potential to result in spinal structure injury and pain. This is consistent with the MRI of five of the seven participants in the pain group, who were reported to have either unilateral or bilateral lumbar stress reactions/fractures at the L5 or L4 level (Table 1). These findings are similar to those reported in other sporting populations, where there is a high rate of stress reactions of the pars interarticularis (30,31).
The potential for the higher lateral flexion forces to be injurious is heightened by the simultaneously occurring peak vertical force (around 10 times the body weight) and lumbar spine extension and right lateral flexion (9,16). Indeed, compression/vertical force combined with extension alone is known to place the fifth lumbar vertebrae under loads that are likely to lead to injury (9). Furthermore, “coupled movements” have been demonstrated to result in greater stress to vertebral structures, including the pars interarticularis, than single plane movements (17,18), and are therefore tentatively linked to the high rates of LBP reported in sporting groups required to repeatedly perform such movements (16). As such, it is proposed that the significantly higher lateral flexion force observed in the current study’s pain group, combined with high vertical forces while the lumbar spine extends and laterally flexes during the drive phase of the tennis serve, is a potential mechanism for tissue strain and LBP in adolescent tennis players. Although there were no significant effects for pain during the forward swing phase, the lumbar region was observed to rapidly flex, left laterally flex, and rotate away from the racquet arm. The lumbar joint reaction moments during this forward swing phase were 3 (sagittal), 8 (transverse,) and 40 (frontal) times larger than those reported during running (34), highlighting the “high” lumbar region loading conditions of the serve and their probable role in LBP development.
Interestingly, while the lumbar region kinematics during the drive phase for both serves was comparable for the participants with and without LBP in the frontal and sagittal planes, the pain group demonstrated significantly less rotation in the drive phase but slightly more in the forward swing phase than in the no pain group. Previous studies have reported that lumbar spine rotation is reduced at the end range of the lumbar spine flexion and extension compared with neutral spine positions (6) and that regional kinematic differences exist in the lumbar spine (25,38), which might explain these discrepancies. That is, it is likely that the extension and lateral flexion of the lumbar spine (in the drive phase) coupled with the increased lateral flexion forces stiffen the spinal segments to limit the extent to which players can rotate. Future investigations should use marker sets that distinguish upper and lower lumbar regions because the lower lumbar spine was the site of stress reaction/fracture among the current cohort and was presented more generally as the region at greatest risk for stress fractures (25,38).
Serve type comparison
The comparison between serve types revealed some differences in lumbar region kinetics and ball position data. In support of previous research, the ball was displaced further from the body in the lateral (leftward) direction during the kick serve (32,35). However, in contrast to the popular belief that the kick serve is associated with greater lumbar region “loading” (35), it was characterized by comparable or slightly lower lumbar region kinetics than the flat serve. Most notably, the flat serve required a larger flexion moment in the forward swing phase, presumably to impact a ball that tended to be displaced further in front of body.
Clinical and coaching implications
We acknowledge that because the pain and no pain comparisons were cross-sectional, it is not possible to distinguish whether these results reflect 1) an adaptation secondary to pain or 2) a mechanism linked with the onset of pain. Therefore, all conclusions should be treated with caution. With this in mind, the finding that tennis players with LBP experience greater left lateral flexion forces when the lumbar spine is under coupled motion and high vertical loading during the serve remains of practical significance. These findings hold implications for trainers, clinicians, and coaches in their attempt to prevent and manage LBP in tennis players.
Increased lateral flexion forces may be associated with different motor control strategies and serve technique in the tennis players with LBP. Future work should investigate whether these findings are present before the development of LBP or represent a either an adaptive or a maladaptive response to LBP and also whether combined coaching and movement training can reduce these potentially injurious forces.
In the drive phase of the serve, coaches place great emphasis on lower limb drive, which is coupled with racket trail and effectively sees the lumbar spine “caught in the middle” or the pivot point about which this coupling occurs. It may be that the emphasis on lower limb drive in the serve inadvertently heightens the loading profile of the lumbar spine in the serve—an interaction that should be considered in future research. Lateral flexion of the spine, or “shoulder-over-shoulder rotation” as it is referred to colloquially, occupies a similarly emphatic role in the teaching of the serve according to many coaches. That is, coaches teach players to assume a “trophy” position at the commencement of the drive phase, while subsequent lateral flexion toward the nonracket arm is preferentially emphasized, often at the expense of trunk rotation in the other planes (forward flexion and rotation). It would therefore seem that this pedagogical emphasis on lateral trunk flexion, while seen as being critical to the development of racket velocity and desirable impact heights and which has in part been motivated by the work of Bahamonde (3), may need to be considered in a holistic evidence-based critique of the development of LBP in adolescent tennis players. In surn, intervention studies featuring detailed analysis of lumbar region mechanics are necessary to determine whether loading profiles can be moderated without compromising racquet velocity. Further, given the high repetition of the serving in tennis, as well as the significant growth and development that accompanies the adolescent period, these two factors deserve consideration in future research designs.
This study was limited to the population described. Therefore, the results may not be generalizable to other ages and females. No EMG data were collected, limiting conclusions of specific motor patterns associated with LBP.
This study detailed lumbar kinetics during the tennis flat and kick serve, confirming that both serves are epitomized by high lumbar loading conditions (14,22), although its findings challenge the popular perception that the kick serve creates greater lumbar “loading” (35). The large left lateral flexion forces that coincide with sizeable vertical force and coupled extension/lateral flexion movements during the drive phase are a likely mechanism for LBP in elite, adolescent tennis players. Notwithstanding the limitations of between-group comparisons, these findings provide the platform for more extensive research and initial information for coaches and health practitioners on the technical factors related to LBP in tennis players.
Funding for this study was provided by Tennis Australia and Curtin University.
The authors have no disclosures or conflicts of interest.
This project was undertaken with the support of a Curtin University Linkage Scheme grant, a Tennis Australia’s Sport Science and Medicine Research grant and the Australian Institute of Sport.
The results of the present study do not constitute endorsement by the American College of Sports Medicine.
1. Adams MA, McMillan DW, Green TP, Dolan P. Sustained loading generates stress concentrations in lumbar intervertebral discs. Spine (Phila Pa 1976)
. 1996; 21 (4): 434–8.
2. Alyas F, Turner M, Connell D. MRI findings in the lumbar spines of asymptomatic, adolescent, elite tennis players. Br J Sports Med
. 2007; 41 (11): 836–41; discussion 41.
3. Bahamonde RE. Changes in angular momentum during the tennis serve. J Sports Sci
. 2000; 18 (8): 579–92.
4. Balague F, Dudler J, Nordin M. Low-back pain in children. Lancet
. 2003; 361 (9367): 1403–4.
5. Barile A, Limbucci N, Splendiani A, Gallucci M, Masciocchi C. Spinal injury in sport. Eur J Radiol
. 2007; 62 (1): 68–78.
6. Burnett A, O’Sullivan P, Ankarberg L, et al. Lower lumbar spine axial rotation is reduced in end-range sagittal postures when compared to a neutral spine posture. Manual Ther
. 2008; 13 (4): 300–6.
7. Campbell A, Alderson J, Lloyd DG, Elliott B. Effects of different technical coordinate system definitions on the three dimensional representation of the glenohumeral joint centre. Med Biol Eng Comput
. 2009; 47 (5): 543–50.
8. Campbell AC, Lloyd DG, Alderson JA, Elliott BC. MRI development and validation of two new predictive methods of glenohumeral joint centre location identification and comparison with established techniques. J Biomech
. 2009; 42 (10): 1527–32.
9. Chosa E, Totoribe K, Tajima N. A biomechanical study of lumbar spondylolysis based on a three-dimensional finite element method. J Orthop Res
. 2004; 22 (1): 158–63.
10. Chow JW, Shim JH, Lim YT. Lower trunk muscle activity during the tennis serve. J Sci Med Sport
. 2003; 6 (4): 512–8.
11. Crewe H, Campbell A, Elliott B, Alderson J. Kinetic sensitivity of a new lumbo-pelvic model to variation in segment parameter input. J Appl Biomech
. In press.
12. de Leva P. Adjustments to Zatsiorsky–Seluyanov’s segment inertia parameters. J Biomech
. 1996; 29 (9): 1223–30.
13. Ferdinands RE, Kersting U, Marshall RN. Three-dimensional lumbar segment kinetics of fast bowling in cricket. J Biomech
. 2009; 42 (11): 1616–21.
14. Fleisig G, Nicholls R, Elliott B, Escamilla R. Kinematics used by world class tennis players to produce high-velocity serves. Sports Biomech
. 2003; 2 (1): 51–64.
15. Foster D, John D, Elliott B, Ackland T, Fitch K. Back injuries to fast bowlers in cricket: a prospective study. Br J Sports Med
. 1989; 23 (3): 150–4.
16. Glazier PS. Is the ‘crunch factor’ an important consideration in the aetiology of lumbar spine pathology in cricket fast bowlers? Sports Med
. 2010; 40 (10): 809–15.
17. Gunzburg R, Hutton WC, Crane G, Fraser RD. Role of the capsulo-ligamentous structures in rotation and combined flexion–rotation of the lumbar spine. J Spinal Dis
. 1992; 5 (1): 1–7.
18. Haberl H, Cripton PA, Orr TE, et al. Kinematic response of lumbar functional spinal units to axial torsion with and without superimposed compression and flexion/extension. Eur Spine J
. 2004; 13 (6): 560–6.
19. Hjelm N, Werner S, Renstrom P. Injury profile in junior tennis players: a prospective two year study. Knee Surg Sports Traumatol Arthrosc
. 2010; 18 (6): 845–50.
20. Iino Y, Kojima T. Validity of the top-down approach of inverse dynamics analysis in fast and large rotational trunk movements. J Appl Biomech
. 2012; 28 (4): 420–30.
21. Johnson CD, McHugh MP, Wood T, Kibler B. Performance demands of professional male tennis players. Br J Sports Med
. 2006; 40 (8): 696–9; discussion 9.
22. Kibler BW. The 4000 watt tennis player: power development for tennis. Med Sci Tennis
. 2009; 14 (1): 5–8.
23. Kibler WB, Safran MR. Musculoskeletal injuries in the young tennis player. Clin Sports Med
. 2000; 19 (4): 781–92.
24. Maquirriain J, Ghisi JP. The incidence and distribution of stress fractures in elite tennis players. Br J Sports Med
. 2006; 40: 454–9.
25. Mitchell T, O’Sullivan PB, Burnett AF, Straker L, Smith A. Regional differences in lumbar spinal posture and the influence of low back pain. BMC Musculoskelet Disord
. 2008; 9: 152.
26. Mitchinson L, Campbell A, Oldmeadow D, Gibson W, Hopper D. Comparison of upper arm kinematics during a volleyball spike between players with and without a history of shoulder injury. J Appl Biomech
. 2013; 29 (2): 155–64.
27. Pearsall DJ, Reid JG, Livingston LA. Segmental inertial parameters of the human trunk as determined from computed tomography. Ann Biomed Eng
. 1996; 24 (2): 198–210.
28. Pearsall DJ, Reid JG, Ross R. Inertial properties of the human trunk of males determined from magnetic resonance imaging. Ann Biomed Eng
. 1994; 22 (6): 692–706.
29. Pluim BM, Staal JB, Windler GE, Jayanthi N. Tennis injuries: occurrence, aetiology, and prevention. Br. J Sports Med
2006; 40: 415–23.
30. Ranson CA, Burnett AF, King M, Patel N, O’Sullivan PB. The relationship between bowling action classification and three-dimensional lower trunk motion in fast bowlers in cricket. J Sports Sci
. 2008; 26 (3): 267–76.
31. Ranson CA, Kerslake RW, Burnett AF, Batt ME, Abdi S. Magnetic resonance imaging of the lumbar spine in asymptomatic professional fast bowlers in cricket. J Bone Joint Surg Br
. 2005; 87 (8): 1111–6.
32. Reid M, Whiteside D, Elliott B. Serving to different locations: set-up, toss, and racket kinematics of the professional tennis serve. Sports Biomech
. 2011; 10 (4): 407–14.
33. Reid MM, Campbell AC, Elliott BC. Comparison of endpoint data treatment methods for estimation of kinematics and kinetics near impact during the tennis serve. J Appl Biomech
. 2012; 28 (1): 93–8.
34. Seay J, Selbie WS, Hamill J. In vivo
lumbo-sacral forces and moments during constant speed running at different stride lengths. J Sports Sci
. 2008; 26 (14): 1519–29.
35. Sheets AL, Abrams GD, Corazza S, Safran MR, Andriacchi TP. Kinematics differences between the flat, kick, and slice serves measured using a markerless motion capture method. Ann Biomed Eng
. 2011; 39 (12): 3011–20.
36. Standaert CJ. Low back pain in the adolescent athlete. Phys Med Rehabil Clin N Am
. 2008; 19 (2): 287–304, ix.
37. Sward L, Eriksson B, Peterson L. Anthropometric characteristics, passive hip flexion, and spinal mobility in relation to back pain in athletes. Spine
. 1990; 15 (5): 376–82.
38. Wade M, Campbell A, Smith A, Norcott J, O’Sullivan P. Investigation of spinal posture signatures and ground reaction forces during landing in elite female gymnasts. J Appl Biomech
. 2012; 28 (6): 677–86.
39. Woltering HJ. A fortran package for generalized, cross-validatory spline smoothing and differentiation. Adv Engng Softw
. 1986; 8 (2): 104–13.
40. Wu G, Siegler S, Allard P, et al. ISB recommendation on definitions of joint coordinate system of various joints for the reporting of human joint motion. Part I: Ankle, hip and spine. J Biomech
. 2002; 35: 543–8.
41. Wu G, van der Helm FCT, Veeger HEJ, et al. ISB recommendation on definitions of joint coordinate systems of various joints for the reporting of human joint motion. Part II: Shoulder, elbow, wrist and hand. J Biomech
. 2005; 38: 981–92.