An Analysis Based on Subclassiﬁcation

SPINE Volume 35, Number 14, pp 1387–1395©2010, Lippincott Williams & Wilkins

Sitting Postures and Trunk Muscle Activity inAdolescents With and Without Nonspecific ChronicLow Back PainAn Analysis Based on Subclassification

Roslyn G. Astfalck, PT,* Peter B. O’Sullivan, PhD,*† Leon M. Straker, PhD,*†Anne J. Smith, PhD,*† Angus Burnett, PhD,‡ Joao Paulo Caneiro, PT,*and Wim Dankaerts, PhD§¶

Study Design. A preliminary cross-sectional compara-tive study of adolescents with nonspecific chronic lowback pain (NSCLBP) and healthy controls.

Objective. To investigate whether differences in spinalkinematic and trunk muscle activity exist in both usualand slump sitting in adolescents with NSCLBP.

Summary of Background. Evidence suggests that lowback pain commonly develops in adolescence and in-creases the risk for low back pain in adulthood. Sitting isan important consideration in adolescents with NSCLBP:currently there are no reports investigating their motorcontrol strategies in sitting.

Methods. Twenty-eight adolescents (14 female) withNSCLBP and 28 matched pain-free controls were re-cruited from a large cohort study. Pain subjects weresubclassified based on O’Sullivan’s classification system.Three-dimensional lumbo-pelvic kinematic data and theactivation of 3 back and 2 abdominal muscles were re-corded during usual and slump sitting. The flexion-relax-ation phenomenon in sitting was also investigated.

Results. Spinal posture in usual and slump sittingwere similar for adolescents with and without NSCLBP.However, differences were identified in both sitting con-ditions when those with NSCLPB were subclassified andcompared with controls. Muscle activation differences werenot consistently identified, with only lower levels of internaloblique activation in usual sitting in NSCLBP compared

with pain-free controls showing significance. Flexion re-laxation was observed in both iliocostalis and thoracicerector spinae in the NSCLBP group but not controls.

Conclusion. This study provides preliminary results.Differences with sitting posture are only seen when ado-lescents with NSCLBP are classified. Trunk muscle acti-vation is not a sensitive marker for discriminating sub-groups of NSCLBP during adolescence.

Key words: chronic low back pain, adolescence, sub-groups, sitting, flexion-relaxation, EMG, kinematics,Raine study. Spine 2010;35:1387–1395

Low back pain (LBP) in adolescence has high prevalence1

and recurrence rates,2,3 that increase with age,3–6 and isassociated with the recurrence of LBP through adultlife.7–9 For some, LBP can be transient and trivial, yet forothers it is chronic and disabling. In adolescence, preva-lence rates for chronic LBP (CLBP) are documented at8%,10,11 with the majority of these disorders classified asnonspecific CLBP (NSCLBP).12,13 Investigating NSCLBPin adolescents may provide insight into a disorder thatcommonly presents in adulthood.

In adults with LBP, sitting is a common aggravatingfactor14,15 and accounts for significant disability.12,16,17

It is reported that adolescents spend large portions oftime in sitting, and those who spend more time flexed, orslumped, report more thoracolumbar pain.15,18–20 A re-cent study of NSCLBP in adolescents found nearly all(92.9%) reported sitting as the most prevalent aggravat-ing factor and contributes most to disability.21

There is growing evidence that adult NSCLBP is not ahomogenous group, but rather represents a series of sub-groups based on both physical and psychosocial vari-ables.12 A mechanism-based classification system hasbeen proposed by O’Sullivan22 which subgroups patientswith localized CLBP based on pain provocative spinalpostures and movement patterns.12,23 Altered spinalpostures and trunk muscle activity during sitting havebeen reported in adults with NSCLBP when subclassifiedbased on this classification system.17 To date there is noevidence as to whether these subgroups exist in adoles-cent NSCLBP populations.

The flexion relaxation phenomenon (FRP) is the pres-ence of myoelectric silence of the back extensors thatoccurs at end range spinal flexion when moving fromstanding to forward bending.24–26 This has also been

From the *Curtin University of Technology, Perth, Western Australia;†Telethon Institute of Child Health Research, Perth Western Australia;‡Edith Cowan University, Perth, Western Australia; §MusculoskeletalUnit, Department of Rehabilitation Sciences, Faculty of Kinesiology andRehabilitation Sciences K. U. Leuven, Leuven, Belgium; ¶Department ofHealth Care, University College Limburg, AUHL-PHL, REVAL–Rehabilitation and Health Care Research Center, Hasselt, Belgium.Acknowledgment date: May 17, 2009. Revision date: July 28, 2009.Acceptance date: August 20, 2009.The manuscript submitted does not contain information about medicaldevice(s)/drug(s).Funds were received in support of this work. No benefits in any formhave been or will be received from a commercial party related directlyor indirectly to the subject of this manuscript.Supported by a National Health and Medical Research Council ofAustralia Project grant (#323200) (to L. M. S. and A. J. S.), the Tele-thon Institute for Child Health Research, the Raine Medical Founda-tion at the University of Western Australia, the Arthritis Foundation ofAustralia and the Arthritis Foundation of Western Australia.The authors A. B. and W. D. were previously associated with the CurtinUniversity of Technology, Perth, Western Australia.This research was approved by the Human Research Ethics Commit-tees of Curtin University of Technology and Princess Margaret Hospi-tal, Perth, Western Australia.Address correspondence and reprint requests to Peter O’Sullivan, PhD,School of Physiotherapy, Curtin University of Technology, GPO BoxU1987, PERTH WA 6845, Australia; E-mail: [email protected]

1387

demonstrated in lumbar multifidus in adults when mov-ing from upright to slump sitting,16,27 but is absent in asubgroup with NSCLBP.16 No studies have investigatedwhether the different subgroups and associated motorcontrol changes identified in adults with NSCLBP arepresent in adolescents with the disorder.

The overall aim of this article was to investigatewhether spinal kinematic and trunk muscle activity dif-ferences exist in both usual and slump sitting in adoles-cents with NSCLBP (when considered as a whole andwhen subclassified) compared with a no-LBP group.

Materials and Methods

Participants were taken from the Western Australian Preg-nancy Cohort (Raine) Study, a longitudinal cohort of over2800 children born in a maternity hospital in Perth, Australia(www.rainestudy.org.au). All adolescents with previously doc-umented NSCLBP were identified from data collected on thecohort during 2003–2005 and were screened and recruited byphone interview (in 2006) to ensure each met participationcriteria (Table 1). Of those suitable for this study, 28 (14 fe-male) volunteered. A comparable control group were identifiedfrom the 2003–2005 data based on gender, age (�6 months),pubertal stage,28,29 and socio-economic status (Australian Bu-reau of Statistics Index of Relative Socio-Economic Advantage/Disadvantage index) (�1 quartile).30 Measures were repeatedduring data collection for this study in one laboratory session atthe School of Physiotherapy, Curtin University of Technologythrough 2006–2007. Controls did not differ from pain subjectson BMI, pubertal stage, and socio-economic status, but wereslightly older, mean difference 3 months (P � 0.049) (Table 2).Prior adult studies showed 2 groups of N � 28 had sufficientpower (�90%) to identify differences in spinal angles17 and trunkmuscle activation16 in both usual and slump sitting of half of one

standard deviation in these measures. Ethics approval was ob-tained and parents provided written informed consent.

For the NSCLBP group, pain intensity was measured with aVisual Analogue Scale, which has high test-retest reliability andgood construct validity when used by adolescents.31 Disabilitywas assessed with the Oswestry Disability Questionnaire32;which was modified for use in minors by removing the questionconcerning sex-life. The questionnaire was measured out of 45for the remaining 9 questions and converted to a percentage–representing percentage disability. The Oswestry is reliable andvalid for use in adults32 and has previously been used in ado-lescents.33 Fear avoidance behavior was quantified by theTampa Scale of Kinesiophobia, the reliability and validity areconsidered moderate-good34–36 and has previously been usedwith adolescents.33

Subjective Clinical AssessmentDuring initial interview each subject with pain was asked astandard series of questions to describe the history and behav-ior of their CLBP (aggravating and easing postures and move-ments and medication usage).

Functional Movement AssessmentAt time of data collection a video recording from a single cam-era was taken by the primary researcher of each subject withNSCLBP during functional postures and movements (usual sit-ting, slump sitting, return to usual sitting, sitting to standing,standing, left and right single-leg-stance, forward bending, andbackward bending). This sequence was completed twice torecord movements from both posterior and postero-lateralviews as described by Dankaerts et al.23

SubclassificationThe pain behavior and video data were used to subclassifysubjects as described by Dankaerts et al23 by a specialist mus-culoskeletal physiotherapist independent of data collection, seeTable 3.

Sitting Posture and Muscle Activity. Three-dimensionallumbo-pelvic kinematic data and trunk muscle activity were

Table 1. Inclusion and Exclusion Criteria at Recruitment

Inclusion Criteria Exclusion Criteria

Male or female Specific diagnosis associated with LBPsuch as spondylolisthesis, discprolapse, inflammatory disorders

Age: 14–16 yr Presence of other conditions affectingthe spine including neurological ormetatastic disease

BMI: �28 kg/m2* Any neurological deficitPain group Any surgery involving the lumbar spine

History of NSCLBP�12 wk duration

Any diagnosed pelvic or abdominalpain disorder in the last 12 mo

Pain in the area fromT12 to glutealfolds

Pregnancy or less than 6 mopostpartum

No peripheral painreferral

Any lower limb surgery in the last 2 yr

Moderate ongoing LBP Current lower limb injuryAverage daily pain

level–VAS �3/10An inability to understand written or

spoken EnglishExperienced most

days of the weekInability to assume test postures

Mechanically inducedlocalized LBP

Control group:No history of spinal

pain

*Required for successful superficial EMG recording.

Table 2. Sample Characteristics*

NSCLBP(Pooled)(n � 28)

No-LBP(n � 28)

NSCLBP (Subclassified)

Extension(n � 13)

Flexion(n � 15)

Age (yr) 15.4 (0.5)† 15.7 (0.5)† 15.4 (0.6) 15.4 (0.5)BMI (kg/m2) 22.2 (3.5) 21.2 (2.6) 22.8 (4.2) 21.6 (2.8)Index of relative

socio-economicadvantage/disadvantage

988.9 (59.7) 979.8 (61.9) 984.8 (51.6) 989.1 (69.0)

Developmental stageGenital 3.5 (0.5) 3.4 (0.6) 3.5 (0.5) 3.5 (0.5)Breast (girls only) 3.9 (0.7) 3.8 (0.8) 3.8 (0.6) 4.0 (0.8)

Usual pain (VAS outof 10)

4.4 � 1.9 5.1 � 1.3 3.8 � 2.1

Pain duration (mo) 26.6 � 12 24.8 � 14.7 28.0 � 8.7Kinesiophobia (total

score 68)36.1 � 10.1 35.5 � 4.7 36.7 � 8.8

Disability (%) 17.9 � 10.1 17.6 � 7.9 18.2 � 11.9No. female subjects 14 14 10‡ 4‡No. male subjects 14 14 3‡ 11‡

*All measures were recorded during laboratory testing session in 2006/2007.†tdf � �2.01054, P � 0.049.‡�2 � 7.0362, P � 0.030.

1388 Spine • Volume 35 • Number 14 • 2010

recorded during usual and slump sitting. Subjects sat on aheight adjustable stool (with no back support), hips and kneesat 90°, feet positioned shoulder width apart and arms relaxedby their side. To standardize head posture, subjects focused ona visual target set at eye level 1.5 m directly in front. Two sittingpositions were investigated: usual sitting–where subjects wereasked to “sit on the stool as you would usually sit”; andslumped sitting–achieved by relaxing the thoracolumbar spineand posterior pelvic rotation. Slump sitting was demonstratedby the investigator before testing, with no manual feedbackprovided through the task. Both positions were maintained for5 seconds; with 3 repetitions of usual/slump/usual sitting.

Spinal PostureSpinal and pelvic angles were measured using Fastrak (3-SpaceFastrak, Polhemus Navigation Sciences Division, VT). Sensorswere taped to the skin over the S2, L3, and T12 spinous pro-cesses. From these sacral, lower lumber, upper lumbar, andlumbar angles were obtained (Figure 1). A negative angle indi-cates anterior sacral tilt for sacral angle measures and lumbarlordosis for all other measures. Procedures for collection andprocessing of these data and the angle definitions have beenoutlined elsewhere.17,37

All angles were calculated for both usual and slumped sit-ting postures. The difference between usual and slumped sittingfor each angle was calculated as the slumped sitting spinal angleminus the usual sitting spinal angle and termed the Spinal Flex-ion Angle Index (SFAI). Kinematic data were averaged acrossthe 3 trials for each subject. The intraclass correlation coeffi-cients for kinematic measures ranged from 0.882 to 0.969 withstandard error of measures ranging from 1.0° to 1.7°.

Spinal Muscle ActivitySurface EMG was collected from 3 back (superficial lumbarmultifidus [multifidus]; iliocostalis lumborum pars thoracis [il-iocostalis]; and longissimus thoracis pars lumborum [thoracicerector spinae]) and 2 abdominal muscles (transverse fibers ofabdominal internal oblique [internal oblique]; and abdominalexternal oblique ([external oblique]). Electrode placements and

procedures for skin preparation and data collection are re-ported elsewhere.16,38

Raw EMG data were visually inspected for heartbeat arti-fact and where present it was minimized by using a fourth-order Butterworth high pass filter with a cut-off frequency of 30Hz. Raw EMG data were then demeaned, full wave rectifiedand band pass filtered (4–400 Hz) to generate a linear enve-lope.16,38 EMG data were amplitude normalized using stan-dardized tasks designed to elicit a stable submaximal voluntaryisometric contraction (sub-MVIC). Normalization protocolshave been detailed elsewhere,16 and have been shown to bereliable both within-day and between-days.38 Normalizedmuscle activation for usual and slumped sitting was averagedacross the 3 trials for each subject.

Two methods were used to investigate FRP. A paired t testwas used to determine whether a significant difference in mus-cle activity existed between the 2 sitting postures for each mus-cle within each group. And to allow direct comparison betweengroups, a flexion-relaxation ratio (FRR) was calculated by di-viding the averaged muscle activation in usual sitting by that inslumped sitting.16,39,40

The existence of differences in muscle activation levels betweenleft and right sides were examined by use of paired t tests. Asignificant difference between sides was observed for externaloblique (P � 0.018) only. For muscles other than external obliqueleft and right side data were pooled, whereas, external oblique wasanalyzed bilaterally. The intraclass correlation coefficients foreach EMG-based measure ranged from 0.690 to 0.993 with stan-dard error of measures ranging from 2.6% to 6.0% sub-MVIC.

Statistical Analysis. Independent t tests were used to com-pare the differences in kinematics, SFAI, muscle activation, andthe FRR between no-LBP and NSCLBP (pooled) groups. Anal-ysis of covariance (ANCOVAs), with gender as a covariate,

Table 3. Clinical Features of Adolescent NSCLBPMovement Control Disorders

Flexion pattern Movements and postures involving flexion orthe lumbar spine aggravate NSCLBPsymptoms

Spinal extension relieves painProvocative postures and functional tasks

associated with a flexed lumbar spine(e.g., slump sit and squatting)

Active extension pattern Movements and postures involving extensionof the lumbar spine aggravate NSCLBPsymptoms

Spinal flexion relieves painProvocative postures and functional tasks

associated with hyper extension of lumbarspine (hyperlordotic sitting and standing)

Multidirectional pattern All movement directions (flexion andextension) provoke NSCLBP symptoms

Neutral spinal postures relieve painProvocative postures and functional tasks

associated with either flexed or hyper-extended lumbar spine (e.g., slump sit,squatting, and sway stand)

Subjects for this research were classified into these patterns based on theirdirection of pain provocation.12,13,23

Figure 1. Spinal kinematic variables.

1389Sitting in Adolescent NSCLBP • Astfalck et al

were used to determine differences between NSCLBP sub-groups and no-LBPs–the omnibus test. Post hoc pairwise com-parisons by Least Squares Differences, with gender as a covariate,were used to analyze differences between subgroups. Additionally,comparisons of confidence limits for difference and interpretationof effect sizes were used to qualify differences. Paired t tests wereused to determine differences in muscle activity between sittingpostures. SPSS-V13 for Windows (SPSS Chicago, IL) was used toperform all statistical tests with � � 0.05.

Results

Spinal KinematicsNo differences in sitting spinal posture were observedbetween adolescents with and without NSCLBP (P �0.28). Differences between groups were only apparentwhen individuals with NSCLBP were subclassified. Esti-mates of differences between those with and withoutNSCLBP are at Table 3.

In usual sitting, statistically significant differenceswere shown on subgroup analysis for sacral (ANCOVAP � 0.001), upper lumbar (ANCOVA P � 0.001), andlumbar (ANCOVA P � 0.002) angles. Estimated meandifference between subgroups of NSCLBP and controlsafter adjustment for gender are at Table 3. The confi-dence interval estimates suggest a very likely effect of alarge-sized difference between the extension subgroup

and no-LBP for sacral, upper lumbar and lumbar angles(LSD P � 0.002, P � 0.042, P � 0.005, respectively) andbetween the extension and flexion subgroups (LSD allcomparisons P � 0.001) and for the flexion subgroupand no-LBP for upper angle only (LSD P � 0.011) seeTable 4 and Figure 2; but no strong evidence of a differ-ence between those with flexion-related NSCLBP andno-LBP for sacral or lumbar angle.41,42 No statisticallysignificant differences were shown for lower lumbar an-gle (ANCOVA P � 0.093). However, the estimated dif-ferences adjusted for gender, Table 4, suggest differencesmay exist. The confidence interval estimates suggest alikely effect of a moderate sized difference between theextension subgroup and both the flexion subgroup andthe no-LBP group, but no strong evidence of a differencebetween those with flexion-related NSCLBP and no-LBP.41,42 The direction of these differences demonstratethat adolescents with extension pattern NSCLBP satwith greater anterior pelvic tilt and lumbar lordosis,whereas adolescents with flexion pattern NSCLBP dis-played a kyphotic lumbar spine (Figure 2).

In slump sitting, statistically significant differenceswere shown on subgroup analysis for sacral (ANCOVAP � 0.004), upper lumbar (ANCOVA P � 0.023), andlumbar (ANCOVA P � 0.007) angles. Estimated mean

Table 4. Group and Subgroup Means, Standard Deviation and Mean Difference for Spinal Posture and TrunkMuscle Activity

Spinal Angle Trunk Muscle

Degrees % Sub-MVIC

Sacral Angle

LowerLumbarAngle

UpperLumbarAngle

LumbarAngle Multifidus Iliocostalis

ThoracicErectorSpinae

InternalOblique

LeftExternalOblique

RightExternalOblique

Usual sittingNSCLBP �0.7 � 9.7 �4.4 � 9.0 1.9 � 7.9 �2.6 � 13.8 32.9 � 31.2 36.4 � 33.0 52.1 � 29.6 25.6 � 22.2 31.0 � 25.7 28.6 � 27.0No-LBP 2.0 � 9.2 �2.2 � 7.9 0.9 � 7.5 �1.3 � 11.7 27.2 � 23.1 32.2 � 23.9 59.4 � 28.8 34.4 � 28.8 34.7 � 31.1 22.0 � 18.7Extension �8.2 � 8.1 �8.7 � 7.2 �4.0 � 5.5 �12.8 � 11.5 35.1 � 30.4 48.5 � 29.4 60.6 � 29.2 24.3 � 8.9 27.8 � 13.4 28.3 � 25.2Flexion 5.7 � 5.5 �0.7 � 9.0 6.9 � 6.0 6.2 � 8.7 32.2 � 31.3 48.1 � 41.9 50.8 � 28.9 25.9 � 20.3 30.2 � 31.0 19.6 � 18.0No-LBP/NSCLBP �2.8 � 5.0 �2.3 � 4.68 0.9 � 4.1 �1.4 � 6.9 5.7 � 14.8 4.2 � 15.4 �7.2 � 15.6 �8.8 � 13.8* �3.7 � 15.3 6.6 � 12.4No-LBP/extension 9.6 � 5.9* 6.5 � 6 5.4 � 5.3* 11.8 � 8.1* �6.1 � 20.8 �15.5 � 23.2 �2.0 � 20.9 12.3 � 19.0 4.4 � 20.4 �2.8 � 14.7No-LBP/flexion �3.8 � 6.1 0.4 � 5.3 �6.3 � 4.8* �5.9 � 7.5 �8.7 � 19.0 �8.5 � 21.3 7.8 � 19.2 14.1 � 17.4 6.8 � 18.7 �0.9 � 13.5Extension/flexion �12.8 � 6.9* �6.0 � 6.8 �1.7 � 6.0* �17.7 � 9.4* �2.6 � 24.1 7.0 � 27.0 9.7 � 24.2 1.8 � 22.1 2.4 � 23.7 1.9 � 17.1

Slump sittingNSCLBP 10.9 � 9.6 1.0 � 9.4 8.1 � 8.9 9.1 � 14.0 33.6 � 30.3 48.3 � 36.0 55.3 � 28.9 25.1 � 15.8 29.1 � 24.1 23.6 � 21.7No-LBP 14.3 � 9.5 2.2 � 9.0 9.1 � 5.9 11.2 � 12.0 26.1 � 23.1 34.7 � 22.0 60.6 � 25.3 39.1 � 30.2 32.8 � 28.5 21.0 � 16.8Extension 4.1 � 8.7 �4.1 � 7.9 3.5 � 5.0 �0.7 � 11.8 35.2 � 31.1 41.4 � 33.7 58.7 � 32.0 23.8 � 9.8 31.1 � 17.2 37.1 � 33.0Flexion 16.7 � 5.9 5.4 � 8.4 12.1 � 9.7 17.6 � 9.7 30.9 � 32.3 32.0 � 32.9 46.4 � 27.1 27.1 � 29.3 30.9 � 31.9 21.2 � 18.6No-LBP/NSCLBP �3.4 � 5.1 �1.2 � 4.9 �0.9 � 4.1 �2.1 � 7.0 7.4 � 14.4 13.6 � 16.0 �5.3 � 25.1 �14.0 � 12.9 �3.7 � 14.1 6.5 � 12.3No-LBP/extension 10.1 � 6.1* 5.6 � 6.0 5.5 � 5.4* 11.1 � 18.3* �4.2 � 21.2 �9.5 � 22.27 �3.6 � 22 9.0 � 19.8 4.3 � 21.9 �7.7 � 16.7No-LBP/flexion �0.8 � 7.2 �0.9 � 5.4 �3.4 � 5.0 �4.2 � 7.6 �6.9 � 19.3 �2.6 � 20.3 12.3 � 20.2 1.9 � 18.2 8.1 � 20.1 �3.1 � 15.3Extension/flexion �10.9 � 7.1* �6.5 � 6.9 �8.9 � 6.3* �15.3 � 9.6* �2.7 � 24.5 7.0 � 25.7 15.9 � 25.5 �7.1 � 23.0 3.8 � 25.4 4.7 � 19.3

Sitting to slump sittingNSCLBP 11.6 � 6.1 5.4 � 5.0 6.3 � 6.8 11.7 � 10.0 1.1 � 0.3 1.1 � 0.2 1.1 � 0.3 1.1 � 1.2 1.0 � 1.0 0.9 � 0.2No-LBP 12.3 � 8.1 4.4 � 5.2 8.1 � 5.3 12.5 � 9.7 0.9 � 0.2 1.0 � 0.2 1.1 � 0.3 1.2 � 0.4 0.9 � 1.0 1.0 � 0.2Extension 12.3 � 6.9 4.6 � 4.6 7.5 � 6.4 12.1 � 10.4 1.0 � 0.2 1.1 � 0.2 1.1 � 0.3 1.1 � 0.2 1.0 � 0.2 0.9 � 0.3Flexion 11.0 � 5.4 6.1 � 5.3 5.2 � 7.1 11.4 � 10.1 1.2 � 0.3 1.1 � 0.2 1.1 � 0.3 1.1 � 0.4 1.0 � 0.1 1.0 � 0.2No-LBP/NSCLBP �0.7 � 3.8 1.0 � 2.7 �1.8 � 3.2 �0.8 � 5.2 0.14 � 0.12* 0.11 � 0.12 0.05 � 0.14 �0.13 � 0.20 2E-3 � 0.078† �0.07 � 0.12No-LBP/extension 0.5 � 5.6 �0.7 � 4.0 0.1 � 4.6 �0.7 � 7.7 �0.10 � 0.16 �0.10 � 0.15 0.02 � 0.21 0.13 � 0.27 1E-3 � 0.11 0.09 � 0.15No-LBP/flexion 2.4 � 5.1 �1.3 � 3.6 2.9 � 4.4 1.7 � 7.1 �0.17 � 0.15 �0.09 � 0.14 �0.06 � 0.20 0.18 � 0.25 �3E-3 � 0.11 0.07 � 0.14Extension/flexion 1.9 � 6.5 �0.6 � 4.6 2.8 � 5.5 2.4 � 9.0 �0.07 � 0.19 0.02 � 0.18 �0.08 � 0.24 0.05 � 0.31 �4E-3 � 0.03 �0.02 � 0.18

All group or subgroup values are mean � SD.All comparisons are mean difference and 95% CI for difference adjusted for gender.For kinematics, the sitting to slump sitting comparison is the difference sitting–slump sitting.For muscle activity, the sitting to slump sitting comparison is the FRP ratio sitting/slump sitting.*Denotes statistically significant comparison at P � 0.05.†The nomenclature E-3 denotes �10�3 such that e.g. 2E-3 � 2 � 10�3.


difference between subgroups of NSCLBP and controlsafter adjustment for gender are at Table 3. The confi-dence interval estimates suggest a very likely effect of alarge sized differences between the extension subgroupand no-LBP for sacral, upper lumbar and lumbar angles(LSD P � 0.002, P � 0.048, P � 0.009, respectively) andbetween the extension and flexion subgroups (LSD P �0.003, P � 0.007, P � 0.002, respectively); but no strongevidence of a difference between the flexion subgroupand no-LBP for sacral, upper lumbar or lumbar an-gle.41,42 No statistically significant differences wereshown for lower lumbar angle during slump sitting(ANCOVA P � 0.123). However, the estimated differ-ences adjusted for gender, Table 4, suggest differencesmay exist. The confidence interval estimates suggest alikely effect of a moderate sized difference between theextension subgroup and both the flexion subgroup andthe no-LBP group, but no strong evidence of a differencebetween the flexion subgroup and no-LBP.41,42 The direc-tion of these results indicates that the extension subgroupduring slump sitting, sat in more anterior pelvic tilt, lowerlumbar lordosis, and less lumbar kyphosis when comparedto both other groups (Table 4, Figure 3). While the flexiongroup displayed trends towards greater posterior pelvic tiltand lumbar kyphosis than other groups.

There were no differences between no-LBP and sub-groups for SFAI (Figure 4), indicating each group movedthrough similar ranges of motion for each measure.

Trunk Muscle ActivityDuring usual sitting the only difference observed in trunkmuscle activity was a significantly greater activation ofinternal oblique in the no-LBP as compared to theNSCLBP (t � �2.170, P � 0.034) (Figure 5), the confi-dence interval estimates indicate that the likely effect isfor a small size difference. No differences were observed

when individuals were subgrouped. No differences werenoted in muscle activation during slumped sitting be-tween no-LBP and NSCLBP groups or subgroups ofNSCLBP (Figure 6). Significant reductions in muscle ac-tivity between upright and slump sitting were shown foriliocostalis (t � �2.132, P � 0.042) and thoracic erectorspinae (t � �2.128, P � 0.043) for the pain group andfor iliocostalis (t � �2.333, P � 0.038) for the extensionsubgroup suggesting that flexion relaxation was presentin the pain subjects. A significant increase was shown formultifidus (t � 2.760, P � 0.010) for controls, Table 5.Differences were found between no-LBP and NSCLBP

Figure 2. Usual sitting posture. Error bars represent 95% CI. Figure 3. Slump sitting posture. Error bars represent 95% CI.

Figure 4. Spinal flexion angle index. Error bars represent 95% CI.


for multifidus FRR (t � 2.397, P � 0.020), with theconfidence interval estimates indicating that the likelyeffect is for a moderate size difference between groups.The direction of these results indicates less relaxation ofmultifidus in slump sitting for healthy controls than forthose with NSCLBP. While not statistically significant(t � �1.91, P � 0.061), the confidence interval estimatesfor difference suggest a likely effect of a moderate sizedincrease (less relaxation) between no-LBP and NSCLBPgroups for Iliocostalis. No statistically significant differ-ences were shown for the subgroup comparison for mul-tifidus (ANCOVA P � 0.070). However, the estimateddifferences adjusted for gender suggest differences mayexist, Table 4. The confidence interval estimates suggesta likely effect of a moderate sized decrease (more relax-ation) between the no-LBP and flexion subgroup (LSDP � 0.027), but no strong evidence of a difference be-tween the extension subgroup and either the LBP groupor flexion subgroup.41,42 No other muscles displayed dif-ferences in flexion-relaxation (Figure 7).

Discussion

No differences in spinal posture were observed in eitherusual or slump sitting between adolescents with and

without NSCLBP before subgrouping. These results aresimilar to adults for usual sitting, but not for slump sit-ting where differences in adults have been identified be-tween those with and without NSCLBP.17 The levels oftrunk muscle activation in adolescents were variable andlargely nondiscriminatory between groups. In both sittingconditions, only internal oblique was different. In our com-parable adult study, no EMG differences were shown inusual sitting between no-LBP and NSCLBP groups.

Similar to our adult research,16,17 kinematic differ-ences were only observed once the pain group was sub-classified based on clinical presentation into extensionand flexion subgroups. Postural subtypes have also been

Figure 5. Usual sitting (back and abdominal muscles). Error barsrepresent 95% CI.

Figure 6. Slump sitting (back and abdominal muscles). Error barsrepresent 95% CI.

Table 5. Difference in Muscle Activity Between Uprightand Slump Sitting

Group orSubgroup Multifidus Iliocostalis

ThoracicErectorSpinae

InternalOblique

LeftExternalOblique

RightExternalOblique

ControlMean 1.04 �2.52 �1.25 �4.74 1.99 0.99SD 2.00 14.44 11.63 13.42 6.82 4.54t(27)* 2.76 �0.92 �0.57 �1.87 1.54 1.16P 0.01 0.364 0.574 0.073 0.134 0.258

PainMean �0.65 �11.89 �3.22 0.44 1.95 4.97SD 2.72 29.51 8.00 18.74 7.97 15.64t(27)* �1.27 �2.13 �2.13 0.13 1.3 1.68P 0.214 0.042 0.043 0.902 0.206 0.104

ExtensionMean 0.06 �7.1 �1.86 �0.44 3.35 8.82SD 2.76 10.97 7.41 6.14 10.79 21.33t(12)* 0.07 �2.33 �0.9 �0.26 1.12 1.5P 0.944 0.038 0.384 0.8 0.285 0.162

FlexionMean �1.27 �16.04 �4.4 1.21 0.74 1.64SD 2.76 11 7.41 6.14 10.79 21.33t(14)* �0.87 �1.59 �2 0.18 0.65 0.85P 0.082 0.135 0.066 0.856 0.524 0.411

Means and SD are difference in activity. All values are % subMax.A negative value indicates presence of FRP.Bolded values indicate statistically significant differences exist between up-right and slump sitting.*The subscript value in parentheses is the df value.

Figure 7. Flexion relaxation ratio (back and abdominal muscles).Error bars represent 95% CI.


documented in adolescents during standing,43 with non-neutral postures being associated with higher odds forLBP and strong associations between subtype and gen-der.43 This reinforces the importance of subgroupingsubjects with NSCLBP.

Extension PatternThe sitting posture of the extension subgroup was char-acterized by anterior pelvic tilt and a lumbar lordosis. Allkinematic variables discriminated the extension sub-group from no-LBP and the flexion subgroup, except forthe lower lumbar angle, where differences were in a sim-ilar direction but not statistically different at � � 0.05.These results, although similar to comparable adult re-search, differ regionally in that the posture of the lowerlumbar spine was found most discriminatory betweengroups in adults,17 whereas in this study of adolescentsupper lumbar angle was most discriminatory. These dif-ferences may reflect group or maturation differences. Hy-perlordotic postures in adults are associated with in-creased compressive forces on posterior spinal elements,notably the facet joints,44–46 and hyperlordotic posturesare related to LBP in adolescents,43 suggesting that thesepostures may be provocative of LBP.

Muscle activation in the extension subgroup was nodifferent to any other group. This finding is at odds withour adult studies showing greater activation of multifi-dus and iliocostalis in the extension subgroup comparedto pain-free subjects.16 These findings may represent theutilization of spinal postural muscles not measured inthis study (such as iliopsoas or deep back muscles), tomaintain hyperlordotic sitting.

Significant gender differences were noted with clinicalclassification of NSCLBP subjects, similar to our adultstudy findings.17 These differences may reflect inherentgender responses to pain, or relate to different social,cultural or body image influences in this group.

Flexion PatternLinks between assuming end range flexed postures andincreased prevalence of LBP have been reported in bothadolescents15,18,21,33 and adults.37,47 In usual sitting theflexion group displayed a kyphotic lumbar spine com-pared to the extension subgroup, with sacral, upper lum-bar and lumbar angles discriminating between sub-groups. Differences were observed between the flexionand control group on the basis of upper lumbar flexiononly. This is different to our adult research where theflexion subgroup was able to be differentiated from theextension and no-LBP groups by sacral and lower lum-bar angles, which may reflect age based developmentaldifferences. That more males in this study were classifiedas flexion pattern concords with other sitting and stand-ing postural studies in adolescents that show malespresent with less lordotic postures and greater thoracickyphosis than females.43,48–52 It has been suggested thata flexed thoracolumbar posture may increase spinalloading representing a potential mechanism for LBPprovocation.53–56

Levels of muscle activation in the flexion subgroupwere not different to any other group. This is at oddswith our adult study that demonstrated lower muscleactivation in the flexion compared to extension sub-group.16 Slump sitting data suggests that, while ky-photic, adolescents with flexed postures may not be atend of range flexion and therefore retain some degree ofactive muscle tension during sitting.

Usual to Slump SittingIn this research, each group moved through a similarrange of motion from usual to slump sitting, althoughdifferences between groups were observed in the startand end positions. In our comparable adult study, pa-tients with NSCLBP showed less ability to changelumbo-pelvic posture from usual sitting, particularlythrough the lower lumbar spine.17 Differences in resultsmay reflect a greater plasticity in the motor system ofadolescents with NSCLBP.

Flexion-relaxation was evident in iliocostalis and tho-racic erector spinae for the NSCLBP group, and in ilio-costalis for the extension subgroup but not controls. An-tithetically, controls showed a significant increase in theactivation levels of multifidus during slumping. The sizeof this increase was small (1.04% sub-MVIC) and ofquestionable clinical significance. The FRR discrimi-nated between those with and without NSCLBP only formultifidus, with healthy controls exhibiting relaxation ofmultifidus in slump sitting less often than those withNSCLBP. Although greater flexion-relaxation in theflexion subgroup was observed, it was not statisticallysignificant at � � 0.05. This is at odds with previousadult data showing greater relaxation in multifidus andiliocostalis in healthy controls than LBP groups.16 Inadults the FRP is also known to be consistent and repeat-able: visual analysis of the adolescent data showed ahighly variable FRP response across pain and controlgroups, between subjects and within subject trials. Thesefindings may reflect an immature spinal motor controlsystem, where the loss of flexion-relaxation observed inadult CLBP is not a feature of adolescent CLBP.

Clinical ImplicationsSimilar to adults, subgroups of adolescents withNSCLBP can be identified clinically and confirmed withpostural analysis in usual and slump sitting. The direc-tion of postural differences, with extension subgroupsexhibiting hyperlordotic sitting postures and flexion sub-groups kyphotic sitting postures, are similar to those ob-served in adults.17,47,57 The magnitude of the differencesin spinal posture are able to be detected clinically ashighlighted by the fact that these subgroups were identi-fied visually from video footage. The irony is that, aswith previous adult studies, those who reported painprovocation with extension activities and postures satmore extended. Subjects who reported pain provocationwith flexion activities and postures sat more flexed thanthe extension group. This research highlights the poten-tial importance of identifying subgroups of NSCLBP in


the examination, clinical management and scientific in-vestigation of the disorder. This is supported by recentresearch where a short physiotherapy intervention re-sulted both in a change in sitting posture and a reductionin LBP in adolescents with NSCLBP.58

Unlike adults, muscle activation and FRP were lessdiscriminatory between those with and without pain andsubgroups of NSCLBP. These findings may reflectgreater immaturity and plasticity of the spinal motorcontrol system in adolescents. Alternatively, it may re-flect increased levels of pain, disability and duration ofLBP in the adult group where FRP is absent.16

Research LimitationsIt is acknowledged that due to the small sample size theresults presented here should be considered as prelimi-nary. Further, the small sample size may have reducedthe power of the study to detect differences in muscleactivation, although comparable sample sizes in adultshave identified differences. While the inclusion and ex-clusion criteria may also restrict the generalisability ofthese results, the consistency of postural results withadult data supports the validity and generalisability ofthese results. Due to the short duration of the sittingtasks, the effects of fatigue on sitting motor control werenot considered. Future investigations could include longexposure seated tasks, control for anthropometrical factorssuch as lumbar spine height and be powered sufficiently todetermine interaction of effects of gender and NSCLBP sub-group on sitting posture. The preliminary findings pre-sented here, and a comparison to our previous work inadults raise some interesting observations but require fur-ther confirmation by a larger investigation.

Conclusion

While the study is preliminary, the following conclusionscan be made:

1. Sitting spinal posture did not discriminate adoles-cent NSCLBP from pain free controls unless ado-lescents with NSCLBP were subgrouped.

2. Adolescents in flexion and extension pain subgroupswere different from pain free controls on the basis ofspinal kinematics, similar to findings in adults.

3. Trunk muscle activation was not a sensitive marker fordiscriminating subgroups of NSCLBP in adolescents.

4. Flexion-relaxation in sitting was evident in iliocos-talis and thoracic erector spinae for adolescentswith NSCLPB but not for healthy controls.

Key Points

● Spinal kinematics does not discriminate adolescentNSCLBP from pain free controls unless subclassified.

● Subgroups of adolescent NSCLBP can be identi-fied on the basis of spinal kinematics.

● Trunk muscle activation is not a sensitive markerfor discriminating subgroups of adolescentNSCLBP.

● Flexion relaxation phenomenon in sitting wasevident in iliocostalis and thoracic erector spinaefor adolescents with NSCLPB but not for healthycontrols.

AcknowledgmentsThe authors are extremely grateful to all the participantsand families who took part in this study and the wholeRaine Study team, which includes data collectors, cohortmanagers, data managers, clerical staff, research scien-tists, and volunteers. The authors thank Jemma Colemanand Paul Davey, research assistants, Curtin University ofTechnology.

References

1. Ebrall PS. The epidemiology of male adolescent low back pain in a northsuburban population of Melbourne, Australia. J Manipulative Physiol Ther1994;17:447–53.

2. Burton AK, Clarke RD, McClune TD, et al. The natural history of low backpain in adolescents. Spine 1996;21:2323–8.

3. Taimela S, Kujala UM, Salminen JJ, et al. The prevalence of low back painamong children and adolescents. A nationwide, cohort-based questionnairesurvey in Finland. Spine 1997;22:1132–6.

4. Balague F, Nordin M, Skovron ML, et al. Non-specific low-back pain amongschoolchildren: a field survey with analysis of some associated factors. J Spi-nal Disord 1994;7:374–9.

5. Leboeuf-Yde C, Kyvik KO. At what age does low back pain become a com-mon problem? A study of 29,424 individuals aged 12–41 years. Spine 1998;23:228–34.

6. Watson KD, Papageorgiou AC, Jones GT, et al. Low back pain in school-children: occurrence and characteristics. Pain 2002;97:87–92.

7. Brattberg G. Do pain problems in young school children persist into earlyadulthood? A 13-year follow-up. Eur J Pain 2004;8:187–99.

8. Harreby M, Neergaard K, Hesselsoe G, et al. Are radiologic changes in thethoracic and lumbar spine of adolescents risk factors for low back pain inadults? A 25-year prospective cohort study of 640 school children. Spine1995;20:2298–302.

9. Kopec JA, Sayre EC. Stressful experiences in childhood and chronic backpain in the general population. Clin J Pain 2005;21:478–83.

10. Bejia I, Abid N, Ben Salem K, et al. Low back pain in a cohort of 622Tunisian schoolchildren and adolescents: an epidemiological study. EurSpine J 2005;14:331–6.

11. Salminen JJ, Erkintalo MO, Pentti J, et al. Recurrent low back pain and earlydisc degeneration in the young. Spine 1999;24:1316–21.

12. O’Sullivan P. Diagnosis and classification of chronic low back pain disor-ders: maladaptive movement and motor control impairments as underlyingmechanism. Man Ther 2005;10:242–55.

13. O’Sullivan PB. ‘Clincal instability’ of the lumbar spine: its pathological basis,diagnosis and conservative management. In: Jull GA, Boyling JD, eds.Grieve’s Modern Manual Therapy: The Vertebral Column. 3rd ed. Edin-burgh, United Kingdom: Churchill Livingstone; 2004:311–31.

14. Balague F, Troussier B, Salminen JJ. Non-specific low back pain in childrenand adolescents: risk factors. Eur Spine J 1999;8:429–38.

15. Geldhof E, De Clercq D, De Bourdeaudhuij I, et al. Classroom postures of8–12 year old children. Ergonomics 2007;50:1571–81.

16. Dankaerts W, O’Sullivan P, Burnett A, et al. Altered patterns of superficialtrunk muscle activation during sitting in nonspecific chronic low back painpatients: importance of subclassification. Spine 2006;31:2017–23.

17. Dankaerts W, O’Sullivan P, Burnett A, et al. Differences in sitting posturesare associated with nonspecific chronic low back pain disorders when pa-tients are subclassified. Spine 2006;31:698–704.

18. Murphy S, Buckle P, Stubbs D. Classroom posture and self-reported backand neck pain in schoolchildren. Appl Ergon 2004;35:113–20.


19. Salminen JJ. The adolescent back. A field survey of 370 Finnish schoolchil-dren. Acta Paediatr Scand Suppl 1984;315:1–122.

20. Sjolie AN. Associations between activities and low back pain in adolescents.Scand J Med Sci Sports 2004;14:352–9.

21. Astfalck RG, O’Sullivan PB, Straker L, et al. Characteristics of non-specificchronic low back pain in an adolescent population. Paper presented at:Musculoskeletal Physiotherapy Australia 15th Biennial Conference; 2007;Cairns, QLD.

22. O’Sullivan PB. Lumbar segmental ‘instability’: clinical presentation and spe-cific stabilizing exercise management. Man Ther 2000;5:2–12.

23. Dankaerts W, O’Sullivan PB, Straker LM, et al. The inter-examiner reliabil-ity of a classification method for non-specific chronic low back pain patientswith motor control impairment. Man Ther 2006;11:28–39.

24. Kaigle AM, Wessberg P, Hansson TH. Muscular and kinematic behavior ofthe lumbar spine during flexion-extension. J Spinal Disord 1998;11:163–74.

25. McGorry RW, Hsiang SM, Fathallah FA, et al. Timing of activation of theerector spinae and hamstrings during a trunk flexion and extension task.Spine 2001;26:418–25.

26. Sihvonen T, Partanen J, Hanninen O, et al. Electric behavior of low backmuscles during lumbar pelvic rhythm in low back pain patients and healthycontrols. Arch Phys Med Rehabil 1991;72:1080–7.

27. O’Sullivan P, Dankaerts W, Burnett A, et al. Evaluation of the flexion relax-ation phenomenon of the trunk muscles in sitting. Spine 2006;31:2009–16.

28. Tanner JM. Normal growth and techniques of growth assessment. Clin En-docrinol Metab 1986;15:411–51.

29. Tanner JM, Whitehouse RH. Clinical longitudinal standards for height,weight, height velocity, weight velocity, and stages of puberty. Arch DisChild 1976;51:170–9.

30. Trewin D. 2039.0—Information Paper—Census of Population and Hous-ing, Socio-Economic Indexes for Areas—Australia. Canberra, Australia:Australian Bureau of Statistics; 2001.

31. Stinson JN, Kavanagh T, Yamada J, et al. Systematic review of the psycho-metric properties, interpretability and feasibility of self-report pain intensitymeasures for use in clinical trials in children and adolescents. Pain 2006;125:143–57.

32. Fairbank JC, Couper J, Davies JB, et al. The Oswestry low back pain dis-ability questionnaire. Physiotherapy 1980;66:271–3.

33. Perich D, Burnett A, O’Sullivan PB. Low back pain and the factors associatedwith it: examination of adolescent female rowers. In: Schwameder H, Srtu-tzenberger G, Fastenvauer V, et al, eds. XXIVth Symposium on Biomechan-ics in Sports. Salzburg, Austria: The University of Salszburg; 2006:355–8.

34. Goubert L, Crombez G, Van Damme S, et al. Confirmatory factor analysis ofthe Tampa Scale for Kinesiophobia: invariant two-factor model across lowback pain patients and fibromyalgia patients. Clin J Pain 2004;20:103–10.

35. Swinkels-Meewisse EJ, Swinkels RA, Verbeek AL, et al. Psychometric prop-erties of the Tampa Scale for kinesiophobia and the fear-avoidance beliefsquestionnaire in acute low back pain. Man Ther 2003;8:29–36.

36. Vlaeyen JW, Kole-Snijders AM, Boeren RG, et al. Fear of movement/(re)injury in chronic low back pain and its relation to behavioral perfor-mance. Pain 1995;62:363–72.

37. Burnett AF, Cornelius MW, Dankaerts W, et al. Spinal kinematics and trunkmuscle activity in cyclists: a comparison between healthy controls and non-specific chronic low back pain subjects-a pilot investigation. Man Ther 2004;9:211–19.

38. Dankaerts W, O’Sullivan PB, Burnett AF, et al. Reliability of EMG measure-ments for trunk muscles during maximal and sub-maximal voluntary iso-

metric contractions in healthy controls and CLBP patients. J ElectromyogrKinesiol 2004;14:333–42.

39. Burnett A, O’Sullivan P, Caneiro JP, et al. An examination of the flexion-relaxation phenomenon in the cervical spine in lumbo-pelvic sitting. J Elec-tromyogr Kinesiol 2009;19:e229–36.

40. Watson PJ, Booker CK, Main CJ, et al. Surface electromyography in theidentification of chronic low back pain patients: the development of theflexion relaxation ratio. Clin Biomech (Bristol, Avon) 1997;12:165–71.

41. Hopkins WG. A spreadsheet for deriving a confidence interval, mechanisticinference and clinical inference from a P value. Sportscience 2007;11:16–20.

42. Hopkins WG, Marshall SW, Batterham AM, et al. Progressive statistics forstudies in sports medicine and exercise science. Med Sci Sports Exerc 2009;41:3–13.

43. Smith A, O’Sullivan P, Straker L. Classification of sagittal thoraco-lumbo-pelvic alignment of the adolescent spine in standing and its relationship tolow back pain. Spine 2008;33:2101–7.

44. Adams MA, Dolan P. Spine biomechanics. J Biomech 2005;38:1972–83.45. Haberl H, Cripton PA, Orr TE, et al. Kinematic response of lumbar func-

tional spinal units to axial torsion with and without superimposed compres-sion and flexion/extension. Eur Spine J 2004;13:560–6.

46. Schendel MJ, Wood KB, Buttermann GR, et al. Experimental measurementof ligament force, facet force, and segment motion in the human lumbarspine. J Biomech 1993;26:427–38.

47. O’Sullivan PB, Mitchell T, Bulich P, et al. The relationship beween postureand back muscle endurance in industrial workers with flexion-related lowback pain. Man Ther 2006;11:264–71.

48. Dunk NM, Callaghan JP. Gender-based differences in postural responses toseated exposures. Clin Biomech (Bristol, Avon) 2005;20:1101–10.

49. Poussa MS, Heliovaara MM, Seitsamo JT, et al. Development of spinalposture in a cohort of children from the age of 11 to 22 years. Eur Spine J2005;14:738–42.

50. Straker LM, O’Sullivan PB, Smith AJ, et al. Relationships between prolongedneck/shoulder pain and sitting spinal posture in male and female adolescents.Man Ther 2009;14:321–29.

51. Straker LM, O’Sullivan PB, Smith AJ, et al. Sitting spinal posture in adoles-cents differs between genders, but is not clearly related to neck/shoulder pain:an observational study. Aust J Physiother 2008;54:127–33.

52. Widhe T. Spine: posture, mobility and pain. A longitudinal study from child-hood to adolescence. Eur Spine J 2001;10:118–23.

53. Callaghan JP, McGill SM. Low back joint loading and kinematics duringstanding and unsupported sitting. Ergonomics 2001;44:280–94.

54. Callaghan JP, McGill SM. Intervertebral disc herniation: studies on a porcinemodel exposed to highly repetitive flexion/extension motion with compres-sive force. Clin Biomech (Bristol, Avon) 2001;16:28–37.

55. Granata KP, Wilson SE. Trunk posture and spinal stability. Clin Biomech(Bristol, Avon) 2001;16:650–9.

56. Scannell JP, McGill SM. Lumbar posture–should it, and can it, be modified?A study of passive tissue stiffness and lumbar position during activities ofdaily living. Phys Ther 2003;83:907–17.

57. Rose SJ. Physical therapy diagnosis: role and function. Phys Ther 1989;69:535–7.

58. Perich D, Burnett A, O’Sullivan PB, et al. Low back pain in adolescent femalerowers: a multi-disciplinary intervention study. In: Huiskes R, Guilak F, eds.XXIst Congress of the International Society of Biomechanics. Taipei, Tai-wan: International Society of Biomechanics; 2007:S654.


An Analysis Based on Subclassiﬁcation

Documents

Transcript of An Analysis Based on Subclassiﬁcation