Printer Friendly

Risk factors for lower back injury in male dancers performing ballet lifts.

The hours of physically demanding work required for ballet training often results in dancers succumbing to injury at some stage in their careers. High incidences of overuse (1,2) and lower back injuries (LBI) have been reported in dance epidemiology. (3-6) A LBI can lead to extended time away from dance, or even premature retirement. The likelihood of male dancers developing LBIs has been suggested to be higher than for female dancers, due to the lifting tasks they perform in ballet choreography. (5-10) Previous scientific publications concerning lifting in dance have been descriptive (11) or performance based, (12) and therefore very limited data are available to suggest which variables, if any, influence male dancers' risk of LBI while performing ballet lifts.

Lower back injuries have been established as one of the most frequently occurring occupational injuries, and workers with lifting or manual handling duties have been identified as having greater LBI risk. (13-16) These facts have led to the generation of methods for quantifying the potential for LBIs while performing various tasks. The 3D Static Strength Prediction Program (3DSSPP, University of Michigan, Centre for Ergonomics, 2006, Version 5.0.6) was developed as a means of estimating forces about the lumbar spine during quasi-static manual handling tasks, and has been used to analyze a range of occupational tasks, from publicans lifting a keg of beer to paramedics lifting a patient. (17,18) The National Institute of Occupational Safety and Health (NIOSH) has published a series of guidelines and compression force limits that define variables associated with the onset of LBI, such as lift rate and duration, hand load magnitudes, and orientation relative to the lifter. (19) Most lumbar force estimation models are limited to static posture analyses, and therefore do not take into account trunk extension velocities that have been linked to the incidence of LBIs in industry. (14,20)

The peak lumbar anterior shear force (PLASF) generated during lifting has been identified as a predictor of LBI. (14,20) Peak lumbar anterior shear force occurs when a person lifts a load situated in front of his or her body with a flexed trunk posture. The weight of the lifter's upper body and the lifted mass create a downward force, or a moment, on the lumbar spine that presses the L5 vertebrae anteriorly and inferiorly onto the L5/ S1 vertebral disc. The upward force generated by the lifter's lower limbs creates a force that pushes the S1 vertebrae superiorly and posteriorly onto the L5/S1 vertebral disc above. The resultant effect is compression and shear forces at the L5/S1 joint that compress the anterior margin of the L5/S1 vertebral disc and stress the L5/S1 facet joints. The lumbar spine has been demonstrated in vitro to have lower resistance to anterior shear force than to compression force. (21-24) Lumbar spinal tissues have also been reported to have lower resistance to combined shear and compression force than to axial compression alone. (25) Subjecting the lumbar spine to combined shear and compression forces, therefore, increases the likelihood of LBI. (26) Thus, it may be that the combination of shear and compression lumbar force generated by male dancers during ballet lifts poses a significant, but as yet unquantified, risk of LBI.

The aim of this 3D biomechanical analysis of ballet lifts was to compare the L5/S1 peak lumbar anterior shear and corresponding compression forces in male dancers performing two lifts. Lumbar compression forces will be referenced to NIOSH normative data as an indicator of male dancer LBI risk. Selected biomechanical variables will be compared between the lifts, and retrospective regression analyses will be used to identify those variables that significantly predict PLASF.


Eight male dancers (mean age 22.6 [+ or -] 6.1 years) and five ballerinas (mean age 26.8 [+ or -] 10.6 years) were recruited from the West Australian Ballet Company (males, n = 5; ballerinas, n = 3), the West Australian Academy of Performing Arts (males, n = 2; ballerina, n = 1), and the Diana Waldron Ballet Academy (male, n = 1; ballerina, n = 1). The same ballerina was not used for all trials, as the West Australian Ballet Company requested that male dancers lift their familiar partners. Height and body mass were measured for both the male dancers and ballerinas. All male dancers were asked to complete a questionnaire outlining their history of back injury. All participants completed a consent form that informed them of their rights, and the study was approved by the University of Western Australia Human Research Ethics Committee.

Prior to the collection of lifting data 54 retro-reflective markers were placed on each male dancer's limbs, torso and head, to meet the requirements of the UWA customised marker set and model. The trajectories of the markers were used to create 3D representations of the participants' bodies. An additional five markers were placed on the male dancers' lumbar region (Fig. 1) to create a lumbar segment, required for the calculation of shear and compression forces at L5/S1. All ballerinas were fitted with six markers on pelvic bony landmarks and two markers on each foot, such that the pelvis segment of the ballerina could be constructed in 3D and the point at which the ballerina left the ground during each lift could be identified.

A 12-camera Vicon MX 3D motion analysis system (Oxford Metrics, Oxford, UK) and one 1200 mm x 1200 mm AMTI force plate (AMTI, Watertown, MA, USA) were used to collect 3D kinematics and kinetics during the lifting trials. Raw ground reaction force (GRF) data were sampled at 2000 Hz, with retro-reflective 3D marker trajecto ry data sampled at 25 0 Hz. During modelling, the Vicon Workstation software automatically re-sampled the force plate data to the same frequency as the 3D motion data. Prior to data collection, cameras and capture volume were calibrated in accordance with Vicon generic procedures. Calibrations that resulted in camera residuals greater than 0.1% of the distance from the camera to the center of the capture volume (generally > 1.0 mm) during any data capture session were recalibrated.


All male dancers completed lifting tasks in two conditions:

1. Full Press (FP): The male dancer lifts the ballerina above his head from the waist, finishing in a pose with the ballerina lying supine on the male dancer's hands with her trunk in hyperextension (Fig. 2). In this lift the ballerina begins in releve, drops quickly into demi-plie, and then jumps vertically into the lift.

2. Arabesque (AR): The male dancer lifts the ballerina, who is in an arabesque position, above his head; one hand is placed on the torso and the other proximal to the knee of the extended leg (Fig. 3). At initiation of this lift the ballerina does not jump; beginning on demi-pointe she is lifted and maintains a static arabesque pose throughout the duration of the lift.

Five trials of each lift were captured for each participant, of which three trials were selected for analysis based on the level of marker occlusion present in the trial. The order of the lifts was randomized across the subject pool, and adequate recovery time was allowed between trials to control for fatigue.


Data Processing

The 3D marker trajectories and force plate data were filtered at a cut-off frequency of 5 Hz using a low pass Butterworth recursive filter, which was determined using residual analysis and visual inspection. The peak lumbar anterior shear force was identified using a customised Bodybuilder (Oxford Metrics, Oxford, UK) 3D dynamic model, with a standard inverse dynamics approach. This required the GRF data to be connected to the mid-pelvis, which was defined as the center of the left and right anterior and posterior superior iliac spines. A lumbar segment was created using the five markers placed on the lumbar region (Fig. 1). The mean point between the LUM1, LUM3, LUM5, LM3L, and LM3R markers determined the lumbar segment origin. This lumbar segment acted as the root segment by which a lower torso segment was joined at an approximated L5/S1 point. This point was calculated using the definition of Zhang and Xiong (27) as 5 cm anterior to the spinous process of the L5 vertebrae (LUM5 marker). The segment and inertial parameters of the pelvis segment were slightly modified to create a lower torso segment consistent with the segment and inertial definitions of de Leva. (25) This segment was defined as encompassing the section from the omphalion (navel) to the mid-point of the hip joint center. The omphalion measurement (vertical distance between the mid- point of the anterior superior iliac spines and the navel) was recorded from the male dancers using a set of callipers as described by de Leva. (28) This approach provided 3D joint reaction forces about the approximated L5/S1 level throughout the duration of the lifts. The direct connection of the GRF to the pelvis will result in an overestimation of the lumbar force values by the 3D motion analysis outputs; however, the primary function of this procedure was the identification of the time of PLASF.


At the time of PLASF, joint angles representing the male dancer's posture were recorded for input into the 3DSSPP model. These absolute angles were calculated with respect to one of the three global coordinate system axes as required by the 3DSSPP model. The 2D position of the pelvis, lower leg and thigh calculated in the 3D dynamic model were input into the 3DSSPP model, which generated a static "stick figure" posture, as shown in Figure 4.

This 3DSSPP model uses a series of inverse dynamic calculations incorporating 2D joint angles, anthropometry data, and kinetic hand loads to calculate external moments. Hand loads (in Newtons) were deemed to be equivalent to the weight of the ballerina, for reasons outlined below. The 3D motion analysis data demonstrated that PLASF for all male dancers occurred at the commencement of the lift, approximately 0.01 seconds prior to the ballerina moving vertically into the lift (Fig. 5). Not surprisingly, this time point was also associated with peak GRF and peak trunk flexion (Figs. 6 and 7). Visual inspection of the ballerina's midpelvis trajectory confirmed that her approximated center of mass was not displacing vertically; hence, any vertical acceleration at the point of peak shear force was small. Therefore, the ballerinas were deemed not to be "assisting" the male dancers by reducing the hand load at this point in the lift. It was further assumed that the ballerina's body weight was evenly distributed between the male dancer's hands, and that no turning moment was generated about the hands during the lifting process. A limitation of the 3DSSPP model is that the lower body kinematic data are only considered in the sagittal plane and the transverse plane of the pelvis is maintained parallel with the horizontal.


All compression forces output from the 3DSSPP were referenced to NIOSH body size and gender-matched normative data as a predictor of potential injury to the lumbar spine. The NIOSH developed the compressive force limits, the Back Compression Design Limit (BCDL), and Back Compression Upper Limit (BCUL) at the L5/S1 joint, to be used as guidelines for lifting within an occupational setting. Any compression force lower than the BCDL of 3400 N indicates a nominal risk to the lifter, whereas compression forces higher than the BCDL require administrative or engineering controls to reduce the potential risk of LBI. A compression force above the BCUL of 6400 N suggests the task poses a serious risk of LBI to the lifter. (19)

Dependant t-tests were performed to ascertain whether the PLASF generated in the two lifting conditions were significantly different for the following independent variables: 3D planar trunk segment angles, trunk extension velocity, participant masses, and male dancer horizontal hand-to-feet distance. Statistical significance was set at p < 0.05, and no Bonferroni correction was applied due to the relatively small number of comparisons made. Retrospective regression analyses were used to identify those variables that significantly predict PLASF.



Height and body mass means of the male dancers were 180.3 cm ([+ or -] 6.88) and 67.15 kg ([+ or -] 6.88), and for the ballerinas were 161.1 cm ([+ or -] 4.61) and 48.15 kg ([+ or -] 5.87), respectively. Four of the eight male dancers reported having been diagnosed with a LBI by a medical professional at some stage in their career; however, all were injury free at the time of testing.

All data samples were considered of equal variance, as skewness and kurtosis absolute levels were less than 2.0. The PLASF for all male dancers occurred at the commencement of the lift, approximately 0.01 seconds prior to the ballerina moving vertically into the lift. Peak lumbar anterior shear force outputs of the 3D motion data and the 3DSSPP were highly correlated. (rxy = 0.868). The compression forces that occurred in synchrony with the PLASF in the FP condition were significantly greater (p < 0.05) than in the AR lift (4726 N vs. 4312 N), while similar PLASFs were recorded across conditions ([approximately equal to]530 N; Table 1).

As the PLASF in the two lifting conditions was not significantly different, the PLAS F data were combined for use in the regression analysis. Descriptive statistics of the independent variables used in the regression analysis are shown in Table 2. The stepwise regression analysis determined peak trunk extension velocity ([beta] = 0.86) and male dancer horizontal hand-to-feet distance (6 = 0.453) as significant (p < 0.05) predictors of PLASF, accounting for greater that 88% of the variance in PLASF ([r.sup.2] = 0.885).


The PLASF for all male dancers occurred at the commencement of the lift, approximately 0.01 seconds prior to the ballerina moving vertically into the lift. This point in the lift coincided with the peak ground reaction force produced by the male dancers (Figs. 6 and 7). Laws and colleagues (12) reported, in a single subject study, that the force between a male dancer's hands and the ballerina was greatest in the lowering phase of lifts similar to those investigated in this study. Comparison with Laws and colleagues' results is difficult due to a difference in reported dependant variables, yet these conflicting results accentuate the need for further investigation in this field.


Mean lumbar compression forces at the point of peak shear for FP and AR lifts (FP 4725 N [+ or -] 852.8 N; AR 4312 N [+ or -] 963.4 N) were above the NIOSH BCDL (3400 N) and below the BCUL (6400 N) limits for a single activity cycle. Forces of these magnitudes indicate a need for administrative controls such as limiting daily lift repetitions in order to reduce the risk of LBI.

Compression forces were significantly higher in the FP than in the AR lift. This could be attributed to a significantly (p < 0.05) larger horizontal distance between the male dancer and the ballerina, or the eccentric loading of the male dancer's lower limb and trunk musculature observed in the FP. In the FP lift the male dancer appears to eccentrically load the lower limb and back extensor lifting muscles at the same time as the ballerina is moving to peak knee flexion, or demi-plie, in preparation for the lift. As primary back extensors attach to the lumbar spine, eccentrically loading the lifting musculature may further increase lumbar compression forces. This dynamic action is a key difference between the FP and AR lifts and may provide an explanation for the difference in compression forces. The results of this study show that PLASF and high compression forces occur prior (~0.01 sec) to the jump of the ballerina; therefore, the ballerina's jump may be of limited value in reducing LBI risk in male dancers. This dynamic coordinated exertion may aid the male dancer once the ballerina has left the ground, but increase the lumbar compression forces at the point of PLASF.


Furthermore, peak trunk extension velocity between the FP and AR conditions was not significantly different, suggesting that the ballerina "jumping into" the lift, as occurs in the FP lift, does not significantly affect peak lifting velocity. This finding indicates that dance practitioners should not assume that the FP lift is less demanding, and by extension less dangerous to a male dancer, as a result of assistance he receives from the ballerina.

The male dancers' horizontal hand-to-feet distances in the FP condition (0.284 [+ or -] 0.076 m) were significantly greater (p < 0.05) than in the AR condition (0.189 [+ or -] 0.054 m). This variable was emphasized by Chaffin and Andersson with reference to the NIOSH guidelines as a strong predictor of lumbar forces and the risk of injury. (19) The male dancer's hand placement on the ballerina for the AR condition is much wider than in the FP and allows the male dancer to lift the ballerina with his hands virtually adjacent to his shoulders (Fig. 3). The larger horizontal distance reported in the FP increases the moment arm of the ballerina's mass on the male dancer's lumbar spine, which would logically increase lumbar joint reaction forces.

The results of the regression analysis demonstrated that the horizontal distance between the male dancer and ballerina could predict PLASF more accurately than the male dancer's 3D planar trunk angles and both the male dancer's and ballerina's body mass. It should be noted that the range of horizontal hand-to-feet distances for the FP and AR lifts was less than 0.1 m. This finding indicates that a relatively small change in the distance between the male dancer and ballerina can have a large effect on male dancer lumbar forces during lifting and becomes more pertinent when considering the precise musical timing involved in ballet, which can provide dancers limited preparation time prior to performing a lift. It is possible, therefore, that a male dancer could be forced to lift a ballerina from an awkwardly increased distance as a result of a performance error. This situation may greatly increase lumbar forces and the risk of LBI.

When considering the low resistance of the L5/S1 joint to combined shear and compression forces, (25,26) the findings of this study indicate a genuine risk of LBI to male dancers during ballet lifts. Of particular concern in this regard are the fatigue-related situations that Liederbach and Compagno found associated with a considerable number of ballet injuries. (29) This suggests that limiting the lift repetitions that are performed each day by male dancers may reduce fatigue, and thereby the potential for inducing an overuse LBI. Yet a paradox exists: if dancers are undertrained they risk increasing lumbar forces during ballet lifts as a result of performance error.


Ballet has received limited biomechanical scientific interest to date, and although acknowledged, many of its athletic requirements have yet to be quantified. Further research into the vast array of ballet movements may facilitate the reduction of often debilitating injuries suffered by dancers, and simultaneously offer insight into the highly skilled motor requirements of ballet performance. Research using advanced lumbar models to investigate the effects of lumbar lordosis on lumbar shear forces would definitely be beneficial. The lumbar force values reported in this study with reference to NIOSH guidelines have demonstrated that ballet lifts pose a LBI risk to the male dancer. The point at which the lumbar forces and risk of LBI are greatest was consistently identified as occurring at the beginning of the lift, before the ballerina left the ground. The distance between the male dancer and ballerina at the beginning of the lift appears to be a highly sensitive predictor of PLASF. Therefore, male dancers need to be well rehearsed with their partners, so as to minimize this distance during performances. A fine balance appears to exist in training male dancers sufficiently to minimize error without overtraining and risking a fatigue-related injury. The highly structured and disciplined organization of ballet companies should allow for a balance of activity and recovery time during rehearsal and performance seasons.


The authors would like to thank Susan Morris for her assistance during the data collection phase of this research.


(1.) Bronner S, Ojofeitimi S, Rose D. Injuries in a modern dance company: effect of comprehensive management on injury incidence and time loss. Am J Sports Med. 2003;31:365-73.

(2.) Solomon R, Micheli LJ, Solomon J, Kelly T. The "cost" of injuries in a professional ballet company: anatomy of a season. Med Probl Perf Art. 1995;10:3-10.

(3.) Coplan JA. Ballet dancers' turnout and it relationship to self reported injury. J Orthop Sports Phys Ther. 2002;32(11):579-84.

(4.) Kadel NJ, Teitz CC, Kronmal RA. Stress fractures in ballet dancers. Am J Sports Med. 1992;20(4):445-9.

(5.) Quirk R. Ballet injuries: the Australian experience. Clin Sports Med.1983;2(3):507-14.

(6.) Bergfeld JA, Hamilton WG, Micheli LJ, et al. Medical problems in ballet: a round table. Phys Sportsmed. 1982;10(3):98-114.

(7.) Ende LS, Wickstrom J. Ballet injuries. Phys Sportsmed. 1982;10(7):100 18.

(8.) Gelabert R. Dancers' spinal syndromes. J Orthop Sports Phys Ther. 1986;7(4):180-91.

(9.) Luke A, Micheli LJ. Management of injuries in the young dancer. J Dance Med Sci. 2000;4(1):6-15.

(10.) Micheli LJ. Back injuries in dancers. Clin Sports Med. 1983;2(3):473-84.

(11.) Lafortune S. A classification of lifts in dance: terminology and biomechanical principles. J Dance Educ. 2008;8(1):13-22

(12.) Laws K, Briel K, Donnalley J, et al. Lifts in partnered dance. Kines Med Dance 1991;13(2):10-21.

(13.) Kelsey JL, White AA. Epidemiology and impact of low-back pain. Spine. 1980;5(2):133-42.

(14.) Marras WS, Lavender SA, Leurgans S E, et al. Biomechanical risk factors for occupationally related low back disorders. Ergonomics. 1995;38(2):377-410.

(15.) Spengler DM, Bigos SJ, Martin NA, et al. Back injuries in industry: a retrospective study. 1. Overview and cost analysis. Spine. 1986;11 (3) :2415.

(16.) Bigos SJ, Spengler DM, Martin NA, et al. Back injuries in industry: a retrospective study. 2. Injury factors. Spine. 1986;11(3):246-51.

(17.) Jones T, Strickfaden M, Kumar S. Physical demands analysis of occupational tasks in neighborhood pubs. Appl Ergon. 2005;36:535-45.

(18.) Lavender SA, Conrad KM, Reichelt PA, et al. Biomechanical analyses of paramedics simulating frequently performed strenuous work tasks. Appl Ergon. 2000;31:167-77.

(19.) Chaffin D B , Andersso n GB. Occupational Biomechanics. New York: Wiley-Interscience, 1984.

(20.) Norman R, Wells R, Neumann P, et al. A comparison of peak vs cumulative physical work exposure risk factors for the reporting of low back pain in the automotive industry. Clin Biomech. 1998;12:561-73.

(21.) Brinckmann P, Biggemann M, Hilweg D. Prediction of the compressive strength of the human lumbar vertebrae. Spine. 1988;14:606-10.

(22.) Cyron BM , Hutton JDG. Spondylolytic fractures. J Bone Joint Surg Br. 1976;58:462-6.

(23.) Miller JAA, Schultz AB, Warwick DN, Spencer DL. Mechanical properties of lumbar spine motion segments under large loads. J Biomech. 1986;19(1):79-84.

(24.) Osvalder AL, Neumann P, Lovsund P, Nordwall A. Ultimate strength of the lumbar spine in flexion--an in vitro study. J Biomech. 1990;23(5):45360.

(25.) Shirazi-adl A, Ahmed AM, Shrivastava SC. Mechanical response of a lumbar motion segment in axial torque alone and combined with compression. Spine. 1986;11(9):914-27.

(26.) Fathallah F A, M arras WS, Parnianpour M. An assessment of complex spinal loads during dynamic lifting tasks. Spine. 1998;23(6):706-16.

(27.) Zhang X, Xiong J. Model-guided derivation of lumbar vertebral kinematics in vivo reveals the difference between external marker-defined and internal segmental rotations. J Biomech. 2003;36:9.

(28.) de Leva P. Adjustments to Zatsiorski- Seluyanov's segment inertia parameters. J Biomech. 1996;29(9):1223-30.

(29.) Liederbach M, Compagno JM. Psychological aspects of fatigue-related injuries in dancers. J Dance Med Sci. 2001;5:116-20.

Jacqueline Alderson, Ph.D., Luke Hopper, B.Sc. (Hons), Bruce Elliott, Ph.D., and Tim Ackland, Ph.D., are from the School of Sport Science, Exercise and Health, The University of Western Australia, Australia.

Correspondence: Jacqueline Alderson, Ph.D., The School of Sport Science, Exercise & Health, The University of Western Australia, 35 Stirling Highway, Crawley, Perth 6009, Australia;
Table 1 Descriptive Statistics for the Shear and Compression 3DSSPP
L5/S1 Force Outputs for the Two Lifting Conditions

Lift Type     Full Press

L5/S1 Force
Component     Shear (N)   Compression (N)

n             8           8
Mean          539.9       4725.8 ([dagger])
SD            55.4        852.8

Lift Type     Arabesque

L5/S1 Force
Component     Shear (N)   Compression (N)

n             8           8
Mean          524.4       4312.0 ([dagger])
SD            58.6        963.4

([dagger]) A significant (p < 0.05) difference was found between the
compression force of the full press and arabesque lifts. * Mean
compression forces for both conditions fell between the NIOSH BCDL
(3400 N) and BCUL (6400 N).

Table 2 The Kinematic and Anthropometric Independent Variables Used in
the Regression Analysis of the Peak Lumbar Shear Forces

                       Trunk Lateral
       Trunk Flexion      Flexion      Trunk Rotation
        ([degrees])     ([degrees])      ([degrees])

Mean        62.6            2.4              4.4
SD           7.9            1.4              3.4

          Peak Trunk                                  Hand-to-Feet
          Velocity       Male Mass   Ballerina Mass     Distance
       ([degrees]/sec)      (kg)          (kg)            (cm)

Mean         99.7           67.1          48.1            23.7
SD           17.5            6.8           5.9             6.6
COPYRIGHT 2009 J. Michael Ryan Publishing Co.
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2009 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Title Annotation:Original Article
Author:Alderson, Jacqueline; Hopper, Luke; Elliott, Bruce; Ackland, Tim
Publication:Journal of Dance Medicine & Science
Date:Jul 1, 2009
Previous Article:Moving beyond the neutral spine: stabilizing the dancer with lumbar extension dysfunction.
Next Article:When can I start pointe work? Guidelines for initiating pointe training.

Terms of use | Privacy policy | Copyright © 2020 Farlex, Inc. | Feedback | For webmasters