Lumbar spondylolysis—a stress fracture of the pars interarticularis of the vertebra—is the most common cause of low back pain in adolescent athletes (21), particularly in sports that involve repetitive flexion–extension of the trunk in combination with rotation (2). One such sporting activity is cricket fast bowling, where an athlete undergoes complex three-dimensional (3-D) trunk motion while experiencing peak vertical ground reaction forces of approximately six times body weight during front foot impact (8). The bowler is required to perform this action with the elbow of the bowling (dominant side) arm relatively straight (with less than 15° of elbow extension permitted between the instant of upper arm horizontal and ball release) and therefore relies to a large extent on trunk and upper arm rotations to build hand speed for maximum velocity at ball release.
In fast bowlers, 80%–90% of spondylolysis have been found to occur on the side of the vertebra contralateral to the dominant arm (9,13,27), suggesting that the asymmetrical nature of the bowling action loads one side of the spine more than the other. Subsequently, a significant amount of research interest surrounding the description of fast bowling mechanics has focused on determining the potential link between technique and injury mechanisms, with the intended outcome being to inform coaching and technique development (12,24). It has been shown using finite element modeling that the stress in the pars interarticularis on one side of the vertebra is highest when a compression force is applied in combination with extension, lateral flexion toward that side, and transverse plane rotation to the opposite side (4), leading bowling biomechanics research to focus on these three trunk motions in relation to potential injury mechanism. Burnett et al. (3) highlighted lumbar lateral flexion toward the contralateral side of the bowling arm as a potential key component in spondylolysis development. This view was based on their findings that bowlers with a mixed action (a technique that involves increased range of upper trunk transverse plane rotation and which has been linked to increased bony injury risk) (8,12) experienced larger lateral flexion angles, range of motion, and angular velocity of the lumbar spine than bowlers with a nonmixed action. Ranson et al. (26) found that fast bowlers exceeded their maximum voluntary lower trunk lateral flexion angle, recorded during an active dynamic range of motion trial, by approximately 30% during the front foot stance phase of the delivery stride. It was further reported that in this position of extreme lateral flexion, the trunk is simultaneously extended and rotated toward the bowling arm side while the bowler is also experiencing high ground reaction forces. This has led authors to postulate that these combined motions, in particular lateral flexion, may be a key element in the etiology of spondylolysis (3,25).
A further potential consequence of the asymmetrical nature of fast bowling is that the quadratus lumborum (QL) muscle—a lateral flexor of the lower trunk—is frequently hypertrophied on the ipsilateral side of the bowling arm (10,25). However, Kountouris et al. (18) described an equal distribution of asymmetry favoring the dominant and nondominant side. The implication of QL asymmetry is disputed as finite element modeling has shown that dominant side QL hypertrophy does not increase the stress in the contralateral pars interarticularis during bowling; however, increased QL asymmetry has previously been associated with low back pain in fast bowlers (15), and Engstrom et al. (10) found a strong association between an enlarged dominant side QL and the incidence of nondominant (contralateral) side spondylolysis. Conversely, a recent prospective study found no relationship between preseason QL asymmetry and spondylolysis incidence during the subsequent cricket season (17). The reason for these contradictory findings is not clear but may be related to differences in methodology or injury risk factors that were not measured, such as bowling technique (8,12,24) and workload (7,22).
No previous studies on QL muscle asymmetry in fast bowlers have been carried out in conjunction with biomechanical analysis. It is suggested that extreme trunk lateral flexion during the delivery stride places severe demands on QL as a mover and a stabilizer (25), but the relationship between the degree of lateral flexion during bowling and the degree of QL asymmetry has not been investigated. The cause of dominant versus nondominant side QL hypertrophy is also unknown. The aim of this study was therefore to analyze lumbo-pelvic lateral flexion kinematics and kinetics in junior fast bowlers and compare bowlers with varying QL cross-sectional area (CSA) asymmetry profiles.
Thirty-nine male asymptomatic fast bowlers (mean age = 16.1 yr, height = 182 cm, mass = 70.8 kg) from Western Australian district and/or state junior cricket squads volunteered to participate in the study. These cricketers are typically involved in fast bowling activity (either in training or in a competitive match) on 3–4 d, bowling approximately 90–200 balls per week. Thirty participants were right-arm bowlers and nine were left-arm bowlers. Ethical approval was obtained from the University of Western Australia’s Human Research Ethics Committee, and all participants (and their guardians, where required) provided informed, written consent to participate in this research. Magnetic resonance imaging (MRI) and 3-D biomechanical bowling analyses were carried out within a 7-d period, before the start of the 2010/2011 cricket season.
MRI assessments were acquired using a GE 1.5-T Signa Excite scanner (General Electric Healthcare, Milwaukee, WI). Axial image slices were orientated to be perpendicular to the L3 vertebral body. T1-weighted sequences with 5-mm slice thickness and 1-mm interslice spacing were performed (repetition time = 500 ms, echo time = minimum, 320 × 224 matrix).
CSA measurements of the left and right QL were taken at the level of the L3–L4 intervertebral disc, which is the level where the CSA of QL is the greatest (14,20). Image quality was rated according to previously described methods (18), and only those images rated 2/3 or 3/3 were accepted for analysis. QL CSA was measured from the MR images by manually tracing the muscle boundaries using a digitization tablet (Intuos2; Wacom Technology Corp., Vancouver, WA) and Mimics software (Version 9.0; Meterialise, Leuven, Belgium). CSA was determined using ImageJ software (National Institutes of Health 2004, http://rsb.info.nih.gov/ij/). Measurements were repeated several months later by the same investigator on images from 10 randomly selected participants, and intraclass correlations were used to examine reliability.
Bowling testing procedure.
Data collection was performed at the biomechanics laboratory at the School of Sport Science, Exercise and Health at the University of Western Australia. A 12-camera VICON MX motion analysis system (Vicon; Oxford Metrics, Oxford, UK) operating at 250 Hz and a 1.2 × 1.2-m force plate (Advanced Mechanical Technology Inc., Watertown, MA) sampling at 2000 Hz were used to collect kinematic and ground reaction force (GRF) data. A cricket crease was marked on the force plate to assist with the collection of isolated front foot kinetics.
To facilitate the capture of positional data of the lower limb, pelvis and lumbar segments for inverse dynamics calculations, retroreflective markers were affixed to the participants’ skin. The customized marker set and model for the lower limbs and pelvis (1) composed of single markers placed on the head of the first and fifth metatarsals, calcaneus, anterior superior iliac spines, and posterior superior iliac spines. A series of marker clusters, consisting of three markers attached to a semirigid plastic baseplate, were attached bilaterally to the thigh and lower legs. Subject-specific static calibration trials were performed using a foot-calibration rig to measure foot abduction/adduction and inversion/eversion angles (1). In addition, static trials were collected with markers placed on the medial and lateral malleoli and medial and lateral femoral condyles and dynamic functional trials performed to determine relevant knee and hip joint axes of rotation and associated joint centers (1). In addition, markers were placed on the L1 and L5 spinous processes and approximately 5 cm on either side of the spine at the level of L4 (LLL and RLL) to define the lumbar segment. A lumbo-pelvic lateral flexion active range of motion (ROM) trial involved the participant laterally flexing the trunk as far as possible with the arms by their sides, keeping both feet flat on the floor and avoiding any rotation or flexion/extension of the trunk. After data processing and analysis, the maximum lumbo-pelvic lateral flexion angle was the output and defined as each individual’s available ROM (26).
After carrying out a self-directed warm-up, participants were required to bowl three overs at match pace. Trials comprising the four highest ball release speeds within these 18 deliveries were selected for analysis. Ball release speed was determined from retroreflective markers attached to the ball and calculated within the motion analysis software.
The 3-D data were processed using Vicon Nexus motion analysis software (Vicon; Oxford Metrics). Data were filtered using a fourth-order low-pass Butterworth filter operating at a cutoff frequency of 15 Hz for the marker trajectories and 50 Hz for the GRF data. Cutoff frequencies were determined using residual analysis (29).
All lower limb anatomical and joint coordinate systems were calculated in accordance with the standards outlined by the International Society of Biomechanics (30) and have been fully described by Besier et al. (1). The lumbar segment was defined using the L5 marker to represent the origin of the lumbar coordinate system. The y-axis was defined using a vector from the L5 to L1 marker, and the x-axis was calculated from the cross product of the y-axis and a defining line between the LLL and RLL markers. The z-axis was calculated as the cross product of the y- and x-axes. The lumbo-pelvic joint location was defined as a virtual point 5% along the length of a line from the L5 marker to the bisector of the two ASIS markers (28) and served as the point of application for the inverse dynamics analysis of the lumbo-pelvic segment. Scaled inertial parameters for the lower limb (5) and pelvis and lumbar segments (23) were incorporated in the inverse dynamics model for the calculation of lumbo-pelvic kinetics.
Lumbo-pelvic kinematics refers to the position of the lumbar segment relative to the pelvis. In addition to calculating the relative angle between lumbar and pelvic segments, the orientation of each these two segments was also calculated relative to the global coordinate system (global segment angles). Although lumbo-pelvic motion occurs in all three planes during bowling, in the current study, we chose to focus on lateral flexion because (i) QL is primarily a lateral flexor of the trunk and, and (ii) bowlers go beyond their available end range in lateral flexion, but not in the sagittal or transverse planes (26). Lateral flexion angles and moments were standardized such that all were reported positive toward the nondominant (contralateral to bowling arm) side. Lumbo-pelvic moments and powers were normalized by the individual’s body mass × height to allow comparison between bowlers of varying height and mass.
Bowlers were categorized into three groups according to QL asymmetry: 1) dominant side QL greater than 10% larger than nondominant side (DOM), 2) nondominant side QL greater than 10% larger than dominant side (NON), and 3) less than 10% difference between the dominant and nondominant side QL (EQUAL) (25).
Statistical analysis was performed using SPSS 19.0 (SPSS Inc., Chicago, IL.). The Kruskal–Wallis test was used for comparison of lumbo-pelvic lateral flexion kinematic and kinetic variables between groups. Where a significant effect was identified, post hoc Mann–Whitney tests were applied to determine where differences occurred. To reduce the likelihood of type I error, a Bonferroni correction was applied and significance set at P < 0.017 (11).
Finally, all subjects with greater than 10% QL asymmetry, regardless of which side was larger, were grouped (UNEQUAL) and compared with the EQUAL group using Mann–Whitney tests, with statistical significance set at P < 0.05.
The intraclass correlation coefficient for repeated measurements of QL CSA was high (0.985, 95% confidence interval = 0.94–0.99). QL asymmetry categorization resulted in 13 subjects (33%) in the EQUAL group, 18 subjects (46%) in the DOM, and 8 subjects (21%) in the NON group (Table 1). Bowlers experienced peak GRFs of 4.8 ± 1.0 body weight (vertical) and 3.3 ± 0.8 body weight (horizontal) and achieved maximum ball release speeds of 30.7 ± 3.5 m·s−1. All participants experienced a large peak lateral flexion moment toward the left, but we observed two peaks within the power data between front foot contact and ball release (the delivery phase) (Fig. 1). For this reason, we chose to examine the peak positive and peak negative lumbar power terms as well as the total positive and negative work in our discrete value analysis (Table 2).
Significant differences in lumbo-pelvic peak lateral flexion angle, angular velocity, moment, positive power, and negative power were observed between the three groups (P < 0.05). Post hoc analysis revealed no significant differences between the DOM and the NON groups. However, the EQUAL group experienced significantly smaller peak angle (18.1° ± 1.9°), moment (10.4 ± 2.8 N·m·kg−1·m−1), positive power (14.7 ± 7.4 W·kg−1·m−1), and negative power (33.8 ± 17.7 W·kg−1·m−1) during the delivery phase compared with the DOM group (angle 22.0° ± 4.5°, moment 13.1 ± 2.3 N·m·kg−1·m−1, positive power 24.2 ± 11.3 W·kg−1·m−1, negative power 51.5 ± 23.0 W·kg−1·m−1; P < 0.017).
When the DOM and the NON groups were combined, bowlers in the UNEQUAL group (n = 26) experienced a larger peak lumbo-pelvic lateral flexion (angle = 21.6° ± 4.2°, angular velocity = 314.9°·s−1 ± 86.6°·s−1, moment = 12.8 ± 2.5 N·m·kg−1·m−1, positive power = 25.6 ± 12.6 W·kg−1·m−1, and negative power = 48.6 ± 20.9 W·kg−1·m−1) compared with the EQUAL group (Table 2) (P < 0.05).
Each bowler exceeded their available active lumbo-pelvic lateral flexion ROM during the delivery phase (mean = 7.8° ± 3.4° beyond available ROM), but there was no difference between the groups.
The present study is the first to examine QL asymmetry in junior fast bowlers in combination with lumbar kinematics and kinetics during bowling. Historically, unilateral QL asymmetry has been thought to occur consistently on the dominant side in fast bowlers (10). However, this has recently been shown to be more variable (18). The distribution of QL CSA profiles in the present cohort is similar to that reported in a group of adult fast bowlers (25), with >10% asymmetry on the dominant side being discovered in almost half of the bowlers. The results of this study therefore support previous research demonstrating that fast bowlers commonly have dominant side QL hypertrophy but that this is not always the case (18,25). The finding that the muscles are also often symmetrical in size (33% of sample) or asymmetrical with a bias to the nondominant side (21% of sample) in this cohort of junior bowlers suggests that varying biomechanical patterns may exist that contribute to the development of different QL asymmetry profiles.
During the delivery phase of the bowling action, bowlers were laterally flexed at the lumbo-pelvic joint toward the nondominant side (Fig. 1). In support of previous research, this population of bowlers exceeded their maximum lateral flexion angle achieved during an active dynamic ROM trial by approximately 8° (26). In addition, the present study reports lumbo-pelvic kinetics and demonstrates that, concurrent to the extreme lateral flexion posture, there was a large lumbar lateral flexion moment toward the nondominant side.
We observed an interesting pattern of lumbo-pelvic lateral flexion motion, with the bowlers typically beginning to laterally flex in the opposite direction (toward the dominant side) after reaching their peak angle near the midpoint of the delivery phase until ball release. Ranson et al. (26) demonstrated a similar pattern. Further examination revealed that the initial increase in lumbo-pelvic lateral flexion angle after front foot contact occurred primarily because of the lateral tilting of the lumbar region, as the pelvis remained relatively stable. Thereafter, there is a rapid lateral tilting of the pelvis toward the nondominant side as the lumbar segment maintains its frontal plane orientation, thus reducing the relative lumbo-pelvic angle (Fig. 2).
The positive power phase that occurs shortly after front foot contact indicates that lumbo-pelvic lateral flexion toward the nondominant side is initially an active motion, implying concentric muscle action of the nondominant side lateral flexors, including QL. In the present cohort, only eight bowlers had enlarged QL on the nondominant side, limiting statistical power when comparing bowlers with different QL asymmetry profiles. However, these bowlers tended to experience larger peak positive power and total positive work than bowlers in the other two categories, which may explain their hypertrophy of QL on this side.
The subsequent lateral tilt of the pelvis coincides with a period of negative power as the lateral flexors of the nondominant side resist the motion by working eccentrically. In this latter half of the delivery phase, the lower trunk muscles of the nondominant side may therefore be working to stabilize the lumbar spine, after initially contributing to active lumbo-pelvic motion. A limitation of inverse dynamics modeling is that the forces and moments produced are resultant values and cannot be apportioned to specific tissues or individual muscles. The current results, while indicating the dominant direction of loading, do not allow us to draw any conclusions on the muscle activity of the dominant side QL; however, the bilateral coactivation of the trunk muscles is known to increase with increased lateral bending velocity (19) and when lifting heavier loads (16).
Fast bowlers with “no” QL asymmetry experienced a significantly smaller peak lumbo-pelvic lateral flexion angle, angular velocity, moment, positive power, and negative power than those with asymmetry of greater than 10% and also tended to experience smaller values on all other lumbo-pelvic lateral flexion variables analyzed (range, angle beyond available active ROM, and total positive and negative work). This suggests that fast bowlers who experience larger lumbo-pelvic lateral flexion loads during bowling may develop greater QL asymmetries. Spondylolysis occurs because of the repetitive loading of the pars interarticularis, and it is therefore logical that the application of greater lumbar loads, while in an extreme position of lateral flexion, increases the risk of lumbar spondylolysis (4). Other authors have reported an association between QL asymmetry and spondylolysis development (10) and low back pain (15) in fast bowlers, and although a recent study determined no relationship between QL asymmetry and spondylolysis incidence, the authors did show that bowlers with >20% asymmetry (50% spondylolysis incidence) were more likely to develop an injury than those with <20% asymmetry (13%) (17). Finite element modeling has demonstrated that the contribution of QL to stresses in the pars interarticularis, when in postures associated with fast bowling, is negligible (6) and thus would be an unlikely cause of spondylolysis. It has also been proposed that the bowling action causes increased bone stress on the nondominant side pars interarticularis and dominant QL hypertrophy develops as an adaptation to reduce these stress levels (6); however, this seems unlikely as QL hypertrophy does not exist exclusively on the dominant side (18). The relationship between QL asymmetry and low back injury (10,15) is therefore most likely incidental, as they are both associated with greater unilateral loading of the spine. This hypothesis has been put forward previously (6,10), but until now, evidence of a relationship between lumbo-pelvic lateral flexion loads and QL asymmetry has been lacking.
Factors such as growth and maturation have not been assessed in the present study, and it is not known whether these may have an influence on asymmetrical QL development.
In conclusion, this study demonstrates that greater lumbo-pelvic lateral flexion loads are associated with QL asymmetry in excess of 10% on either side. The estimation of lumbar kinetics through inverse dynamics modeling requires 3-D motion analysis and is time consuming. MRI screening for QL asymmetry may provide a more feasible and widely available method than 3-D motion analysis to identify bowlers who are experiencing greater unilateral lumbar loads. We propose that QL asymmetry is not a problem in itself and intervention should not focus on reducing muscle imbalance through selective training of the smaller side QL but rather that identification of asymmetry be used as a warning that a bowler may be at increased risk of injury (10) and should be followed up by technique analysis and examination of bowling workload to reduce the likelihood of future injury. A two-dimensional analysis of a bowler’s technique can also be used to assess trunk lateral flexion, although the 3-D nature of the bowling action provides a challenge to camera positioning for frontal plane motion assessment. Future prospective studies are required to investigate whether lateral flexion loading patterns are in fact associated with spondylolysis development. Further, simulation models could be used to determine the contribution of individual muscle forces to further the understanding of lumbar motion and loading during cricket fast bowling.
The authors thank all participants for their involvement and the Perth Radiological Clinic for assisting in performing the MRIs.
The study was supported by a research grant form Cricket Australia.
The authors have no disclosures or conflicts of interest.
The results of the present study do not constitute endorsement by the American College of Sports Medicine.
1. Besier TF, Sturnieks DL, Alderson JA, Lloyd DG. Repeatability of gait data using a functional hip joint centre and a mean helical knee axis. J Biomech
. 2003; 36 (8): 1159–68.
2. Brukner P, Kahn K. Clinical Sports Medicine
. 3rd ed. Sydney: McGraw-Hill; 2007. p. 369–72.
3. Burnett AF, Barrett CJ, Marshall RN, Elliott BC, Day RE. Three-dimensional measurement of lumbar spine kinematics for fast bowlers in cricket. Clin Biomech
. 1998; 13 (8): 574–83.
4. 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.
5. de Leva P. Adjustments to Zatsiorky–Seluyanov’s segment inertia parameters. J Biomech
. 1996; 29 (9): 1223–30.
6. de Visser H, Adam CJ, Crozier S, Pearcy MJ. The role of quadratus lumborum asymmetry in the occurrence of lesions in the lumbar vertebrae of cricket fast bowlers. Med Eng Phys
. 2007; 29 (8): 877–85.
7. Dennis RJ, Finch CF, Farhart PJ, James T, Stretch RA. Is bowling workload a risk factor for injury to Australian junior cricket fast bowlers? Br J Sports Med
. 2005; 39 (11): 843–6.
8. Elliott BC. Back injuries and the fast bowler in cricket. J Sports Sci
. 2000; 18 (12): 983–91.
9. Engstrom CM, Walker DG. Pars interarticularis stress lesions in the lumbar spine of cricket fast bowlers. Med Sci Sports Exerc
. 2007; 39 (1): 28–33.
10. Engstrom CM, Walker DG, Kippers V, Mehnert AJH. Quadratus lumborum asymmetry and L4 pars injury in fast bowlers: a prospective MR study. Med Sci Sports Exerc
. 2007; 39 (6): 910–7.
11. Field A. Discovering Statistics Using SPSS
. 3rd ed. London: Sage Publications; 2009. p. 565–6.
12. 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.
13. Hardcastle P, Annear P, Foster DH, et al.. Spinal abnormalities in young fast bowlers. J Bone Joint Surg Br
. 1992; 74 (3): 421–5.
14. Hides J, Fan T, Stanton W, Stanton P, McMahon K, Wilson S. Psoas and quadratus lumborum muscle asymmetry among elite Australia Football League players. Br J Sports Med
. 2010; 44 (8): 563–7.
15. Hides J, Stanton W, Freke M, Wilson S, McMahon S, Richardson C. MRI study of the size, symmetry and function of the trunk muscles among elite cricketers with and without low back pain. Br J Sports Med
. 2008; 42 (10): 809–13.
16. Huang QM, Andersson EA, Thorstensson A. Specific phase related patterns of trunk muscle activation during lateral lifting and lowering. Acta Physiol Scand
. 2003; 178 (1): 41–50.
17. Kountouris A, Portus M, Cook J. Quadratus lumborum asymmetry and lumbar spine injury in cricket fast bowlers. J Sci Med Sport
. 2012; 15 (5): 393–7.
18. Kountouris A, Portus M, Cook J. Quadratus lumborum asymmetry is not isolated to the dominant side in junior cricket fast bowlers. Br J Sports Med
. 2012; 46 (4): 264–7.
19. Marras WS, Granata KP. Spine loading during trunk lateral bending motions. J Biomechanics
. 1997; 30: 697–703.
20. Marras WS, Jorgensen MJ, Granata KP, Wiand B. Female and male trunk geometry: size and prediction of the spine loading trunk muscles derived from MRI. Clin Biomech
. 2001; 16 (1): 38–46.
21. Micheli L, Wood R. Back pain in young athletes: significant differences from adults in causes and patterns. Arch Pediatr Adolesc Med
. 1995; 149 (1): 15–8.
22. Orchard J, James T, Portus MR, Kountouris A, Dennis R. Fast bowlers in cricket demonstrate up to 3- to 4-week delay between high workloads and increased risk of injury. Am J Sports Med
. 2009; 37 (6): 1186–92.
23. 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.
24. Portus MR, Mason BR, Elliott BC, Pfitzner MC, Done RP. Technique factors related to ball release speed and trunk injuries in high performance cricket fast bowlers. Sports Biomech
. 2004; 3 (2): 263–84.
25. Ranson C, Burnett A, O’Sullivan P, Batt M, Kerslake R. The lumbar paraspinal muscle morphometry of fast bowlers in cricket. Clin J Sport Med
. 2008; 18 (1): 31–7.
26. 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.
27. 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.
28. 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.
29. Winter DA. Biomechanics and Motor Control of Human Movement
. 2nd ed. New York: John Wiley & Sons; 1990. p. 14–44.
30. Wu G, Cavanagh PR. Recommendations for standardization of the reporting of kinematic data. J Biomech
. 1995; 28 (10): 1257–61.