The Effect of Lumbar Posture and Pelvis

Fixation on Back Extensor Torque and

Paravertebral Muscle Activation

Laura J Holder

A thesis submitted to Auckland University of Technology in

partial fulfillment of the requirements for the degree of Master of
Health Science (MHSc)

2013

Faculty of Health and Environmental Sciences

Primary supervisor: Dr Grant A Mawston

 Table of Contents

List of Figures ............................................................................................................. vi

List of Tables.............................................................................................................. vii

Glossary of Abbreviations ........................................................................................ viii

Attestation of Authorship .......................................................................................... ix

Acknowledgements ..................................................................................................... x

Abstract ...................................................................................................................... xi

Chapter 1: Introduction ..................................................................................................... 1

1.1 Overview ............................................................................................................ 1

Chapter 2: Literature Review ............................................................................................. 4

2.1 Introduction ....................................................................................................... 4

2.2 Literature Review Search Process...................................................................... 4

2.3 Epidemiology ..................................................................................................... 5

2.3.1 Prevalence ................................................................................................ 5

2.3.2 Functional anatomy of the paravertebral muscles .................................. 7
2.4 Lumbar Posture................................................................................................ 10

2.4.1 Lumbar posture definitions .................................................................... 10

2.4.2 Posture and back extensor torque ......................................................... 12
2.4.3 Posture and paravertebral muscle activation ........................................ 16
2.4.4 Posture and neuromuscular efficiency ................................................... 19
2.5 Pelvis Fixation .................................................................................................. 22

2.5.1 Pelvis fixation and back extensor torque ............................................... 22

2.5.2 Pelvis fixation and muscle activation ..................................................... 23
2.6 Summary .......................................................................................................... 25

2.7 Aims & Objectives ............................................................................................ 26

2.8 Purpose of the Study ....................................................................................... 26

Chapter 3: Methods ......................................................................................................... 27

3.1 Introduction ..................................................................................................... 27

3.2 Methods ........................................................................................................... 27

ii
 3.2.1 Study design ............................................................................................ 27
3.2.2 Design overview...................................................................................... 27
3.2.3 Participants ............................................................................................. 28
3.2.4 Randomisation process........................................................................... 30
3.3 Experimental Measures ................................................................................... 31

3.3.1 Kinetic data ............................................................................................. 31

3.3.2 Kinematic data ........................................................................................ 32
3.3.2.1 Motion analysis ............................................................................. 32

3.3.2.2 Retro-reflective marker placement............................................... 32

3.3.2.3 Biomechanical modelling .............................................................. 35

3.3.2.4 Axes orientation ............................................................................ 36

3.3.2.5 Extensor moment calculation ....................................................... 36

3.3.2.6 Lumbar posture ............................................................................. 38

3.3.2.7 Lumbar fixation ............................................................................. 39

3.3.3 Electromyography................................................................................... 40
3.3.3.1 Skin preparation ............................................................................ 40

3.3.3.2 Electrodes and placement............................................................. 40

3.3.3.3 EMG collection and processing ..................................................... 40

3.4 Experimental Procedures................................................................................. 42

3.4.1 Equipment calibration ............................................................................ 42

3.4.2 Participant training ................................................................................. 42
3.4.3 Participant preparation .......................................................................... 44
3.4.4 Testing session ........................................................................................ 45
3.5 Statistical Analysis ............................................................................................ 47

Chapter 4: Results ............................................................................................................ 48

4.1 Introduction ..................................................................................................... 48

4.2 Participant Information ................................................................................... 48

4.2.1 Demographic data .................................................................................. 48

4.2.2 Lumbar range of motion ......................................................................... 48
4.3 Data Management ........................................................................................... 49

iii
 4.3.1 Back extensor torque calculation and validation ................................... 49
4.3.2 Torque and gender ................................................................................. 50
4.3.3 EMG and gender ..................................................................................... 52
4.4 Lumbar Posture................................................................................................ 53

4.4.1 Posture and torque/body weight ratio .................................................. 53

4.4.2 Posture and normalised torque .............................................................. 54
4.4.3 Posture and normalised EMG ................................................................. 55
4.4.4 Posture and normalised EMG by muscle division .................................. 56
4.5 Neuromuscular Efficiency ................................................................................ 57

4.6 Pelvis Fixation .................................................................................................. 59

4.6.1 Fixation and normalised torque in each posture ................................... 59

4.6.2 Fixation and normalised EMG for each muscle division......................... 59
Chapter 5: Discussion ...................................................................................................... 61

5.1 Introduction ..................................................................................................... 61

5.2 Lumbar Posture................................................................................................ 61

5.2.1 Posture and torque ................................................................................. 61

5.2.1.1 Gender and torque ........................................................................ 64

5.2.3 Posture and EMG .................................................................................... 66

5.2.4 Posture and EMG for each muscle division ............................................ 69
5.2.5 Neuromuscular efficiency ....................................................................... 71
5.3 Fixation............................................................................................................. 72

5.3.1 Fixation and torque ................................................................................ 72

5.3.2 Fixation and EMG.................................................................................... 73
5.4 Clinical Implications ......................................................................................... 74

5.5 Limitations of the Study ................................................................................... 76

Chapter 6: Conclusion & Recommendations ................................................................... 78

6.1 Conclusion ........................................................................................................ 78

6.2 Recommendations ........................................................................................... 79

Reference List ................................................................................................................... 80

Appendices ....................................................................................................................... 86

Appendix A ................................................................................................................ 86
iv
Appendix B ................................................................................................................ 88

Appendix C ................................................................................................................ 89

Appendix D ................................................................................................................ 93

v
List of Figures Page
2.1 The change in muscle fibre obliquity with change from an upright to a 9
fully flexed posture
2.2 Normalised torque for posture in studies undertaken in standing 15
2.3 Normalised torque for posture in studies undertaken in sitting 15
2.4 The effect of lumbar posture on the mean activation of LES 17
2.5 Neuromuscular efficiency of LES 20
3.1 Recruitment process 30
3.2 Chest harness marker position 31
3.3 Adapted from Visual 3D; view of marker locations 33
3.4 The eight segment, rigid-link biomechanical model (from Visual 3D) 35

3.5 Parameters used to calculate extensor moment 37

3.6 Upper body centre of mass position 38
3.7 Fixation device and set up 39
3.8 Lift initiation postures 42
3.9 Method to establish chain length 45
3.10 Biering-Sorensen position 46
4.1 Mean range of motion of the lumbar spine in standing and in the lift 49
initiation position
4.2 Torque expressed per unit of body weight for each posture and 51
gender
4.3 Torque expressed as a percentage of MVIC during the Biering- 52
Sorensen for each lumbar posture and gender
4.4 Normalised EMG by muscle division for each gender 53
4.5 Torque expressed as a ratio of body weight for each posture 54
4.6 Normalised torque and posture 55
4.7 Normalised EMG for each lifting posture 56
4.8 Mean and SD for normalised EMG for posture and muscle division 57
4.9 The mean neuromuscular efficiency in each posture 58
4.10 NME of each division of muscle in each posture 58
4.11 The effect of pelvis fixation condition in each posture 59
4.12 Normalised EMG and division of muscle comparing fixation 60

vi
List of Tables Page

2.1 Literature search themes and search terms 5

2.2 Definitions of lumbar posture 11
3.1 Pelvis fixation and posture sequence 30
3.2 Anatomical location of retro-reflective markers and body 34
segments formed
4.1 Participant demographics 48
4.2 T-test results comparing the two methods of torque calculation 50
4.3 Mean (SD) peak torque for gender in each test condition 50
4.4 Post hoc analysis of torque/BW between different postures 54
4.5 Post hoc analysis significance levels and CI of normalised EMG 56
4.6 Post hoc analysis of posture on neuromuscular efficiency 57

vii
Glossary of Abbreviations

BE Back Extensor

BW Body Weight

CoM Centre of Mass

CoR Centre of Rotation

EMG Electromyography

ES Erector Spinae

LBI Low Back Injury

LBP Low Back Pain

LES Lower Erector Spinae

MVIC Maximal Voluntary Isometric Contraction

NME Neuro Muscular Efficiency

RMS Root Mean Square

ROM Range Of Movement

UES Upper Erector Spinae

viii
Attestation of Authorship

I hereby declare that this submission is my own work and that, to the best of
my knowledge and belief, it contains no material previously published or written by
another person (except where explicitly defined in the acknowledgements), nor
material which to a substantial extent has been submitted for the award of any other
degree or diploma of a university or other institution of higher learning.

Candidate: ________________________ Date: 12th Nov 2013

ix
Acknowledgements
I would like to thank Grant Mawston for his tireless input into moulding my
(non-existent) academic writing skills and constant support, advice and enthusiasm to
motivate me to complete this challenge. Sorry that I couldn’t always share your
enthusiasm in it being “exciting stuff hey?”, although towards the end it was – a
thousand thank yous.

Thank you to Mark Boocock, for your knowledge and expertise on the
experimental process. While you provided one of the toughest points in this process
by suggesting a change in protocol, it actually provided the most meaningful result of
this research. Yes, you were right … again!

Thanks to Rowan Parsons for your invaluable moral support on student-

Tuesdays, photography skills and fantastic taste in coffee…. it was good fun! Thanks to
Caroline Cross, Bronwyn Harman and Jill Caldwell for your time listening and
supporting me through this process.

Jonathan, I would have missed out on so many opportunities in life if it was not
for you, your encouragement, your belief in me and your ongoing love and support
despite being absent from family life in order to complete this thesis. Thank you to
Janice and Philip for supporting our family to achieve our goals – we love having you
here in New Zealand.

Thanks to Ethan and Jacob, I am soooo looking forward to coming to every

football game, camping holiday, cycle trip and playing with you both and not needing
to justify why I am “sat at that computer doing boring things again”.

Thank you to AUT University for funding my post graduate education.

I would like to acknowledge Auckland University of Technology Ethics

Committee for the ethics approval for this thesis (ethics application number 11/15),
which was granted on 10th May 2011.

x
Abstract
Lumbar postures using a high degree of flexion are considered a risk factor
associated with manual handling related back injuries. The lumbar posture adopted
during lifting is thought to alter the force capabilities of the trunk and the active
contribution of the paravertebral muscles. To date, most studies examining back extensor
(BE) force and muscle activation have been undertaken in non-functional positions with
the pelvis fixated. Furthermore, paravertebral muscle activation has focused on a single
spinal level, despite known anatomical and functional differences of upper erector spinae
(UES), lower erector spinae (LES) and multifidus. An understanding of how lumbar
posture can affect the relationships between trunk extensor torque and paravertebral
muscle activity in functional lifting positions may provide important knowledge for
vocational manual handling training. Thus, the aim of this study was to determine how
lumbar posture and pelvis fixation affect BE torque and muscle activation of 3 divisions of
the paravertebral muscles.

This study involved 26 healthy participants performing a simulated static lift in 3

lumbar postures (lordotic, flexed and mid-range). A maximal isometric lift was undertaken
with and without the pelvis fixated. Participants were provided with real-time feedback of
their lumbar posture using a three dimensional (3D) electromagnetic motion monitor. A
chest harness connected to a 3D floor mounted force gauge via a metal chain, provided a
measure of maximum isometric voluntary BE force. A 9-camera motion analysis system
recorded body position, which was used in conjunction with kinetics, to estimate BE
torque. Muscle activity of the 3 paravertebral muscles (UES, LES, and multifidus) was
recorded using electromyography (EMG). EMG signal for each muscle during peak back
extension in each posture were normalised to that produced during maximal exertion in
the Biering-Sorensen position.

Lumbar posture was found to have a significant effect on BE torque, with torque
increasing from a lordotic to flexed posture (P<0.0001). In contrast, a flexed lumbar
posture produced significantly less muscle activation for all BE muscle groups than a mid
and lordotic posture (P<0.001). As a result, the flexed posture showed a higher

xi
neuromuscular efficiency (NME) ratio (torque/activation) when compared to the mid and
lordotic posture (P<0.001). Each paravertebral muscle responded to a change in posture
in a similar manner. Pelvis fixation was found to have no effect on torque or the intensity
of muscle activation.

Initiating a lift using a lumbar posture similar to the mid position used in this study
may help to avoid loading on the passive structures of the spine, which is thought to occur
when using a fully flexed lumbar posture. The mid posture also offers higher levels of
NME compared to the lordotic lumbar posture, which may be useful in situations when
exerting high BE torque is necessary. Given that pelvis fixation did not influence BE torque
and paravertebral muscle activation, it would suggest that functional lifting tasks should
not be concerned with restraining the pelvis.

xii
 Chapter 1: Introduction

1.1 Overview
A flexed lumbar posture is considered a risk factor associated with low back injury
(LBI) (Hoogendoorn et al., 2000; Punnett, Fine, Keyserling, Herrin, & Chaffin, 1991).
Reducing the risk of LBI during occupations involving manual handling is one of the leading
priorities in preventative medicine and a major concern for health professionals (Waters,
Putz-Anderson, & Garg, 1994). Twenty percent of all injuries in the workplace affect the
back and typically occur during lifting, especially tasks involving high force exertions and
flexed lumbar postures (Hoogendoorn et al., 2000; Waters et al., 1994). Furthermore, LBI
is known to be one of the most common and disabling musculoskeletal conditions, with
the economic cost thought to exceed $500million in New Zealand each year (McBride,
Begg, Herbison, & Buckingham, 2004). The National Institute for Occupational Safety and
Health (NIOSH) have published international guidelines on manual handling (Waters et al.,
1994) in order to minimise the risk of LBI, yet specific parameters regarding lumbar
posture are not given.

Parameters considered important for the prevention of LBI during lifting include
the magnitude of back extensor (BE) torque the spine generates to lift a load, and the
level of paravertebral muscle activation occurring to control the spine (Roy, Keller, &
Colloca, 2003; Tan, Parnianpour, Nordin, Hofer, & Willems, 1993). When generating BE
torque, lumbar posture has been highlighted as an important factor known to influence
the magnitude of the torque, as well as influencing paravertebral muscle recruitment.
Research has shown that when the spine is fully flexed, BE torque is at its greatest, whilst
paravertebral muscle activation levels are at their lowest (Dolan, Mannion, & Adams,
1994). At the end range of lumbar flexion, high loads on spinal structures occur
(Gallagher, Marras, Litsky, & Burr, 2005) and this is when there is minimal muscle
activation to protect the spine, leading to a high risk of LBI when lifting using a flexed
lumbar posture.

1
 The majority of the literature that has investigated the relationship between
lumbar posture and muscle activity during lifting has focused on erector spinae (ES)
muscle activation at one spinal level (usually lower ES at L3) and assumed this to be similar
for all divisions of ES across the lumbar spine (Kaigle, Wessberg, & Hansson, 1998; Roy et
al., 2003; Tan et al., 1993). However, anatomical and EMG studies suggest that the
paravertebral muscles of the lumbar spine have three distinct portions (upper ES, lower ES
and multifidus) that are anatomically and functionally quite different and each can be
identified using EMG (Bojadsen, Silva, Rodrigues, & Amadio, 2000; Ng, Kippers,
Parnianpour, & Richardson, 2002).

The majority of the literature investigating the relationship between lumbar

posture and BE torque while lifting has been undertaken in non-functional positions
where the pelvis and lower limbs have been restrained. The rationale behind using
fixation is to isolate the BE muscles by preventing de-rotation of the pelvis and minimising
the contribution of the pelvis and hip musculature. During lifting the spine does not work
alone but is an important link in the biomechanical chain connecting the upper and lower
body. Less is known about how the BE muscles generate torque and activate in functional
lifting tasks or how pelvis fixation influences the biomechanics of the lumbar spine during
lifting.

The purpose of this study is to investigate the effect how various lumbar postures
during lifting affect BE torque. It also aims to determine muscle activity responses when
lumbar posture changes during lifting in three divisions of the paravertebral muscles. The
final aim is to determine the effect of pelvis fixation on BE torque production and muscle
activation in functional lifting postures to assist in interpreting preceding research which
has used non-functional positions.

The outcome from this study may have benefits for developing manual handling
programmes aimed at reducing the risk of LBI during lifting. It will provide objective
evidence of how the lumbar posture adopted during lifting influences torque and
activation of the BE muscles from a functional lifting perspective. The outcome of this
study will be of significance to health care professionals, health and safety personnel and
2
those regularly involved in manual handling by providing information appropriate to the
prevention and management of LBI. Finally, this study may contribute to the
understanding and implications of lumbar posture during lifting to assist health
professionals educate clients who undertake manual handling tasks in their occupations or
daily activities.

3
 Chapter 2: Literature Review

2.1 Introduction
This chapter begins by outlining how the literature was searched and relevant
articles were included in this review. The relationship between lumbar posture and
epidemiology of LBI is discussed, along with the effect lumbar posture has on the
functional anatomy of the paravertebral muscles. This is followed by discussing how
lumbar posture influences BE torque, paravertebral muscle activation and neuromuscular
efficiency. The effect of pelvis fixation on BE torque and muscle activation are then
discussed. The chapter ends with stating the research aims, objectives and purpose of the
study.

2.2 Literature Review Search Process

A literature search was undertaken in August 2010 and repeated in September
2012. The electronic databases searched were Scopus, Cochrane, Pedro and EBSCOHost
health database, which includes CINAHL, MEDLINE and SPORTdiscus. Searching of the
literature was undertaken in themes to cover different sections of the literature review.
These sections included epidemiology and associated biomechanical risk factors of LBP,
the effect of lumbar posture on torque and muscle activation, and the effect of pelvis
fixation on torque and muscle activation. The search method used different combinations
of search terms to identify potential literature and varied according to the databases,
which are outlined in Table 2.1.

Relevant articles were identified by title and if the main author deemed them
appropriate to this study, the abstract read. Further literature was identified from a
manual search of the in-text citations of these articles. Articles were excluded if they
were not published in English or were related to spinal surgery. Any articles published
before and up until December 2012 were included.

4
 Table 2.1 Literature search themes and search terms

Key search themes

Epidemiology Posture, torque, activation Pelvis Fixation, torque, activation
Risk factor* Lumbar activat* Pelvi* fixation

Epidemiolog* Erector spinae Pelvi* stabilization

Aeiteolog* Back extens* Pelvi*restraint

Prevalence multifidus Muscles

Lifting Electromyograph* Back exercise

Biomechanic* Torque Torque

Kinematic* Moment Activation

Low back Isometric* Electromyograph*

Lumbar Lumbar range Lumbar

2.3 Epidemiology
2.3.1 Prevalence
Low back pain (LBP) affects one in two New Zealanders and is considered one of
the most common, costly and disabling musculoskeletal conditions (McBride et al., 2004).
The cost of LBI to Accident Compensation Corporation (ACC) in 2008/9 exceeded $110
million. In 2008/9 a third of ACC claims were associated with carrying or lifting tasks (ACC,
2010). Also, 80% of all reported injuries sustained in manual handling lifting tasks affected
the lumbar spine (Cole & Grimshaw, 2003).

The high prevalence of LBP has led to a number of epidemiological and

biomechanical studies to identify risk factors associated with LBI in manual handling
occupations. Over 100 risk factors have been identified that relate to the occurrence of
5
LBI in manual handling occupations, but some of the more common mechanical risk
factors include the intensity and frequency of lifting, lifting distance and fatigue (Adams,
Green, & Dolan, 1994; Cole & Grimshaw, 2003; Waters et al., 1994). However, both
epidemiological and biomechanical studies indicate that a key risk factor associated with
LBI is using a lumbar posture with a high degree of flexion (Hoogendoorn et al., 2000;
McGill, Hughson, & Parks, 2000; Punnett et al., 1991).

Most epidemiological studies to date have quantified lumbar posture through the
measurement of trunk flexion by using video analysis or devices capable of measuring
lumbar motion (Hoogendoorn et al., 2000; McGill et al., 2000; Punnett et al., 1991). These
studies have shown that the risk of LBI during lifting is increased when using a lumbar
posture with a high degree of flexion. Punnett et al. (1991) used video recordings to
analyse the lumbar posture of car assembly factory workers. Manual handling tasks
performed in severe lumbar flexion (flexion greater than 45o) were significantly correlated
to a higher incidence of LBI compared to tasks performed in neutral or mild trunk flexion
(0 – 45o of lumbar flexion). Hoogendoorn et al. (2000) also analysed work place video
footage of manual handling lifting tasks and found if flexed postures of more than 60o
were sustained for more than 5% of the working day, there was a high risk factor for LBI.
Marras et al. (1995) assessed 400 workers during their manual handling jobs and
determined lumbar flexion using an tri-axial electrogoniometer worn by the subjects. It
was found that even a small amount of lumbar flexion poses a risk for LBI, but working in
higher degrees of flexion, comes with a higher degree of risk of injury. Marras et al.
concluded that postures that involved lumbar flexion that exceeded 20o were considered
to be a risk factor for LBI.

Biomechanical studies also support epidemiological studies on the negative effects

of using a flexed lumbar posture. Flexed lumbar postures during lifting are thought to
increase the load on the passive systems of the lumbar spine (Bogduk, Macintosh, &
Pearcy, 1992; Dolan et al., 1994) leading to potential injury of the facet joints,
intervertebral discs and/or the posterior ligament complex of the spine (Gallagher &
Marras, 2012; Potvin, McGill, & Norman, 1991). Loading of the passive tissues when

6
moving from the upright neutral spine to approximately 80% of lumbar flexion is
considered minimal, but beyond 80% of available flexion, there is an exponential increase
in the passive tissue resisting the bending moment acting upon the spine (Dolan et al.,
1994). Specifically, it is the posterior ligamentous complex which is considered to exert a
large anterior shear on the spinal segments when end range of flexion is approached
(Gallagher & Marras, 2012). Interestingly, in contrast to shear forces, compressive forces
of the lumbar region have been shown to increases only five percent from an upright
posture to that of full lumbar flexion (Macintosh, Bogduk, & Pearcy, 1993).

2.3.2 Functional anatomy of the paravertebral muscles

The erector spinae (ES) and multifidus are the main BE muscles which exert up to
80% of the extension forces across the lumbar spine (Bogduk, 2012). ES are a muscle
group consisting of several divisions, which have distinct attachments and innervations. In
the lumbar spine, the ES muscle is a complex structure consisting of two main parts;
Longissimus Thoracis and Iliocostalis Lumborum, where the latter is positioned more
lateral than the former (Bogduk, 2012; Macintosh & Bogduk, 1991). Each of these muscles
has two divisions; pars lumborum and pars thoracis, with the thoracic component sitting
more superficially than lumborum. However, clinical examination of the individual
components of erector spinae is not possible because of their close proximity. This has
resulted in a functional classification of the paravertebral muscles according to location of
surface anatomy, and thus they have become commonly referred to as Upper Erector
Spinae (UES) and Lower Erector Spinae (LES), with multifidus being recognised as a
separate muscle (Bogduk et al., 1992; Macintosh et al., 1993; McGill, 1991).

The fascicles of the UES arise from all the thoracic transverse processes and lower
seven ribs merging into the erector spinae aponeurosis, which spans across the lumbar
spine to insert on the sacrum and pelvis (Bogduk, 2012). The ES aponeurosis does not
have any direct attachment onto the lumbar spine. This results in the creation of a
bowstring effect and gives the UES a large mechanical advantage to resist bending
moments (Macintosh & Bogduk, 1991; Macintosh et al., 1993). In the upright posture, the
UES is thought to produce 74% of the total back extensor moment (Bogduk et al., 1992).

7
It has been shown that towards the extreme of lumbar flexion the moment arm of the
UES significantly reduces by up to 39% (Macintosh et al., 1993).

LES attaches to the accessory and transverse processes of each lumbar vertebra
and converges to form the lumbar intermuscular aponeurosis which inserts onto the ilium
near to the posterior superior iliac spine (Bogduk, 2012). The intermuscular aponeurosis
crosses multiple spinal segments but has a shorter lever arm than UES. It is estimated that
in an upright posture, the LES contributes to 15% of total BE torque (Bogduk, 2012).
Lumbar flexion also reduces the ability of the LES to produce extensor torque, although to
a lesser extent than the UES (Macintosh et al., 1993).

Multifidus is the largest, deepest and most medial of the paravertebral muscles. It
consists of short fascicles arranged segmentally, arising from one lumbar vertebra and
descends to attach to one or two lumbar segments caudally, where eventually the lower
lumbar fascicles attach onto the sacrum. Multifidus descends in an oblique orientation,
fanning more lateral as it goes caudally (Bogduk, 2012). The overlying more superficial ES
muscles make clinical examination of multifidus challenging, but below the third lumbar
vertebrae only the ES aponeurosis exist over the top of multifidus making it accessible for
palpation and surface EMG (Bojadsen et al., 2000). It is estimated that 11% of total BE
torque in a neutral lumbar spine originates from multifidus (Bogduk, 2012). Unlike the
UES and LES, an increase in lumbar flexion has a positive effect on increasing the moment
arm of multifidus by up to 30%, which is attributed to the more vertical orientation of the
muscle fibres that span between adjacent segments (Macintosh et al., 1993).

Another key factor that influences back extensor torque of the paravertebral
muscles, is sarcomere length and the position of the overlap of the actin and myosin
filaments (Bogduk, 2012; Brown & Gerling, 2012; Macintosh & Bogduk, 1991; Raschke &
Chaffin, 1996). Flexed lumbar postures elongate the paravertebral muscle fibres and
position the overlap of actin and myosin filaments more optimally for generating extensor
torque (Brown & Gerling, 2012). This is supported by biomechanical studies that found
peak isometric torque capacity and optimal length-tension of the BE muscles to occur at
45o of lumbar flexion (Raschke & Chaffin, 1996). Others consider optimal position for
8
torque to be generated by contractile tissue at up to 80% of lumbar flexion (Dolan et al.,
1994).

While length-tension relationships of the ES muscles may be more optimal near

end range of lumbar flexion, muscle pennation angle and torque generating capacity of
the LES may be reduced in fully flexed lumbar postures. Studies utilising diagnostic
ultrasound have shown that when the spine moves from an upright posture to full flexion,
the pennation angle of the LES reduces from 28o to 10o (McGill et al., 2000; Singh, Bailey,
& Lee, 2011). The change in muscle fascicle obliquity reduces the mechanical advantage
of muscle fibres that generate extensor torque and resist shear forces (McGill et al., 2000)
(Figure 2.1).

Figure 2.1 The change in muscle fibre obliquity with change from an upright (A) to
a fully flexed (B) posture (adapted from Mawston & Boocock (2012))

In summary, while an increase in lumbar flexion improves the ability of the

paravertebral muscles to generate extensor torque because of optimal length-tension and
sarcomere positioning, a reduction in moment arm length and fascicle obliquity of muscle
fibres has the potential to reduce ES muscle extensor torque production and to resist
anterior shear forces towards maximal lumbar flexion. The combination of increased

9
torque with flexion and an inability of muscles to control a rise in shear may present a risk
for LBI.

2.4 Lumbar Posture

2.4.1 Lumbar posture definitions
An array of methods is used to measure lumbar posture, each having its own
classification using different anatomical parameters. Definitions often vary according to
body segments, which can include two or more lumbar vertebrae, incorporation of the
thoracic spine and/or including the sacrum and pelvis (Table 2.2). Trunk inclination can
also be used as an indicator of lumbar posture even though it may predominantly involve
hip flexion and minimal lumbar spine movement. The definitions of lumbar posture
include; lumbosacral angle, pelvisacral angle, trunk inclination, lordotic curvature, lumbar
vertebral angle and lumbar curvature. The definition of these terms is outlined in Table
2.2.

10
 Table 2.2 Definitions of lumbar posture

Author, year Terminology Definition Measurement method

Bogduk, (2012) Lumbosacral angle The wedge shape of the L5/S1 disc and described by the angle of the X-ray
intersection of the line through the inferior surface of L5 with the
line through the upper surface of the sacrum

Lavender, Trafimow, Trunk inclination The motion produced from the lumbar and lower thoracic vertebrae Inclinometer
Andersson, Mayer, & measured by inclining the trunk forward in the sagittal plane
Chen, (1994) measured from T7

During, Goudfrooij, & Lordotic curvature The shape of the lower lumbar vertebral column created by X-ray
Keessen (1985) constructing a circle through the frontal vertices of the vertebral
bodies, defined by three points, L4, L5 and S1

Burgess-Limerick & Lumbar vertebral angle The angle subtended by the lines joining T1, PSIS and the ASIS to the Photographs with
Abernethy (1997) PSIS retro-reflective
markers
During et al. (1985) Pelvisacral angle The angle between the line through the centre of the hip joints and X-ray
the centre of the superior surface of sacrum and the tangent line
through the upper vertices of the sacrum

Roy, Keller, & Colloca Sagittal posture The anterior or posterior inclination of the trunk that requires Isostation,B200
(2003) flexion or extension movement of the lumbar and thoracic spine (isokinetic
without the involvement of pelvic movement machine)

Adams & Dolan (1991); Lumbar curvature: The movement that occurs in the sagittal plane between S1 and L1 Electromagnetic
Dolan & Adams (1993); associated with and/or without pelvic movement Fastrak device
Mannion & Troke (1999)

11
 For the purpose of this thesis, one of the more encompassing definitions for
lumbar posture that incorporates the whole of the lumbar spine was considered to be, the
vertebral movement that occurs in the sagittal plane between the S1 and L1 vertebra
associated with or without pelvic movement (Dolan & Adams, 1993). Vertebral
movement in the sagittal plane produces the physiological movement of flexion and
extension and is the main movement the lumbar spine performs. Concurrent with the
physiological movement of lumbar flexion is forward vertebral translation (Bogduk, 2012).

The amount of sagittal plane movement between L1 and S1 varies greatly within
the literature (Intolo et al., 2009). In a study using electromagnetic sensors (Polhemus
Fastrak device) attached to L1 and S1 on 149 younger healthy adults, lumbar flexion was
performed with participants lower limbs and pelvis secured in a standing frame. Lumbar
range varied from 34o to 75o with a mean of 55o (Dolan et al., 1994). Other studies that
have also measured lumbar curvature between L1 and S1 using a Fastrak lumbar motion
device, where the pelvis and lower limbs were unrestrained, reported a mean range of
lumbar flexion in young healthy adults of 67o (Hindle, Pearcy, Cross, & Miller, 1990) and
71o (Pearcy & Hindle, 1989). A systematic review on lumbar range of motion found a
significant reduction in lumbar flexion with increase in age (Intolo et al., 2009). Gender
differences also exist with females having more range than their male counterparts (Intolo
et al., 2009). It would seem that even when posture is clearly defined and measured using
the same device, the amount of lumbar motion can vary greatly between individuals.

2.4.2 Posture and back extensor torque

Lumbar posture is known to impact upon the ability of the BE muscles to generate
torque. When the lumbar spine becomes more flexed, the paravertebral muscle elongate
and this alter their length-tension and consequently, the torque generating capacity of the
paravertebral muscles (Ward et al., 2009). Several studies have investigated the
relationship between lumbar posture and BE torque and it is consistently shown that
when the spine assumes an increasingly flexed posture the magnitude of BE torque also
increases (Graves et al., 1990b; Holmes et al., 1996; Parnianpour, Campello, &
Sheikhzadeh, 1991; Roy et al., 2003; Smidt et al., 1983; Tan et al., 1993). Conversely,

12
when adopting a more lordotic posture the capacity to generate BE torque diminishes.
The described lumbar posture-torque relationship appears in, males and females (Graves
et al., 1990b; Roy et al., 2003; Smidt et al., 1983), younger participants (Roy et al., 2003;
Tan et al., 1993), older adults (Holmes et al., 1996), healthy individuals (Graves et al.,
1990b; Parnianpour et al., 1991; Roy et al., 2003) and those with LBP (Holmes et al., 1996;
Smidt et al., 1983; Smith, Bissell, Bruce-Low, & Wakefield, 2011). The same finding also
occurs irrespective of method used to measure torque.

There is a gender difference in the torque generating capacity of the BE muscles,

with males being stronger than females (Graves et al., 1990a; Keller & Roy, 2002; Smidt et
al., 1983). Smidt et al. (1983) investigated maximal isometric back extension torque on 12
females and 12 males using a trunk dynamometer in a seated position from 20 o of
extension to 40o of lumbar flexion in 20o increments. Men were found to be significantly
stronger than females, on average by 57%. These findings also concur with Graves and
colleagues (1990a) who undertook a similar protocol using 56 men and 80 women, in
which males were found to be stronger than females in all lumbar postures. These gender
differences still remained when torque was normalised per unit of body weight. Keller &
Roy (2002) also found a gender difference in maximal isometric BE torque, but when
torque was normalised according to body weight and height, no gender differences
existed in torque production. The importance of normalising data in order to compare
genders was further highlighted by Roy et al. (2003) who tested the maximal isometric
torque of 10 male and 10 female participants from 20o of extension to 50o of flexion while
standing with pelvis and lower limbs fixated in a B200 isostation device. When torque in
each lumbar posture was expressed as a percent of torque in the upright standing, no
gender differences existed.

The magnitude of BE torque is not only dependent upon gender but also the test
position. Previous studies tend to have used either standing or sitting positions. In
standing, the BE muscles have been shown to produce maximum torques between 164
Nm and 265 Nm at the end flexion test position in young adults (Roy et al., 2003; Tan et
al., 1993). In contrast, studies using a sitting test procedure have found far greater back

13
extensor torques of between 320 Nm and 345 Nm at end flexion (Graves et al., 1990b;
Smidt et al., 1983). These extensor torque values are based upon 0o of flexion to be the
upright standing or sitting positions. Arguably, when assuming upright sitting, 90 o of hip
flexion will induce posterior rotation of the pelvis which will place the lumbar spine into
more flexion compared to upright standing (Yasukouchi & Isayama, 1995). Therefore,
when testing BE torque in sitting, subjects are more likely to adopt postures where the
lumbar spine is more flexed compared to standing (Callaghan & McGill, 2001). The
alteration in low back joint loading and kinematics during standing compared to
unsupported sitting may partly explain why higher torques are recorded in sitting than
upright standing.

The angle-torque relationship is also noteworthy. Studies performed in standing

showed a linear increase in torque from 0o of flexion to 50o (Parnianpour, Li, Nordin, &
Kahanovitz, 1989; Roy et al., 2003; Tan et al., 1993) (Figure 2.2). In studies testing BE
torque in sitting the torque-angle relationship is more variable between studies with
Graves et al. (1990b) and Holmes et al. (1996) finding only a small increase in torque from
neutral upright sitting to end range of flexion (Figure 2.3). In contrast Smidt et al. (1983)
found the biggest change in torque output in the early range of flexion with similar levels
of torque in 20o and 40o of lumbar flexion. Interestingly, at the end test position in sitting,
torque production appears to reduce in some cases (Graves et al., 1990b; Smidt et al.,
1983).

14
 210

190

Normalised torque (%)

170

Roy
150
Tan
130
Parnianpour
110

90
0 10 20 30 40 50 60
Lumbar flexion (degrees)

Figure 2.2 Normalised torque for posture in studies undertaken in standing

160

150
Normalised torque (%)

140

130
Holmes
120
Graves
110 Smidt

100

90
0 10 20 30 40 50
Lumbar flexion (degrees) Note: Standard deviation
values were not published
for Graves or Smidt

Figure 2.3 Normalised torque for posture in studies undertaken in sitting

There have been few studies that have assessed extensor torque at the extreme of
lumbar flexion. Previous studies have measured torque production at absolute angles of
lumbar flexion up to 50o of lumbar flexion. For example, the end test position used by Tan

15
et al. (1993) and Parnianpour et al. (1989) was 35o, which equates to only 50% percent of
total lumbar flexion (Pearcy & Hindle, 1989). The most flexed test position adopted in a
study was 50o (Roy et al., 2003), which was approximately 70% of maximal lumbar flexion,
according to Pearcy & Hindle (1989). Further research is needed to assess extensor
torque in end range flexion as this posture is often adopted during lifting (Hoogendoorn et
al., 2000) and this is when there is significant recruitment of passive structures of the
spine, which is associated with risk of injury (Dolan et al., 1994; Hoogendoorn et al., 2000).

2.4.3 Posture and paravertebral muscle activation

Lumbar posture is known to influence paravertebral muscle activation (Kaigle et
al., 1998; Kippers & Parker, 1984; Roy et al., 2003; Tan et al., 1993; Toussaint et al., 1995).
Muscle activation is commonly quantified using amplitude measures of surface or fine
wire EMG (Soderberg & Knutson, 2000). Only a small number of studies have examined
activation of the paravertebral muscles during maximal isometric voluntary contraction at
different points of lumbar flexion.

One of the first groups to investigate the effects of lumbar posture on ES muscle
activation was Marras, King & Joynt (1984) who used fine wire EMG to examine levels of
the left and right ES activation during maximal isometric back extension, with minimal
fixation, when standing in neutral (0o), mid (22.5o) and near-end range (45o) lumbar
flexion. EMG data were normalised against the maximum value recorded across all
lumbar postures. Despite not stipulating the exact lumbar spinal level where ES activation
was recorded, activation patterns were similar for left and right ES. ES were found to be
less activated in mid and near end-range postures when compared to upright standing
(neutral posture). However, ES activation levels did not differ significantly between mid
and near-end range postures where it was thought an increase in latissimus dorsi
recruitment, due to its superior mechanical advantage, may have compensated for a
reduction in ES activation with flexion.

Tan et al. (1993) also undertook a similar study to that of Marras et al. (1984),
looking at the effect of lumbar posture on maximum isometric BE torque in three lumbar
postures (0o, 15o and 35o flexed). Thirty-one healthy men were positioned with their
16
pelvis and lower limbs fixated into in a B200 isostation dynamometer with muscle
activation being recorded bilaterally form the medial and lateral ES, level with L3. No
difference in activation levels were found from the medial and lateral portions of ES. In
contrast to Marras and colleagues, Tan et al. found activation levels of ES to significantly
increase from upright standing to 15o flexion. Activation levels continued to rise at a
slower rate thereafter and no significant difference between the mid and end test point
was found (Figure 2.4).

More recently, Roy et al. (2003) examined male and female participants maximum
isometric back extensions while recording activation bilaterally from the LES (level with L3)
using surface EMG. Participants were placed in an isokinetic machine with their pelvis and
lower limbs restrained. Lumbar motion was tested from 20o of extension to 50o of flexion
in 10o increments, with 0o being upright standing. No difference was found between the
activation levels of left and right LES. From 20o of extension to 30o of flexion the
activation levels between males and females were similar. From 40o onwards, men
continued to increase muscle recruitment, while females exhibited a decline in activation.
This gender difference was not statistically significant resulting in a net reduction in
activation of the LES at 50o compared to the peak at 40o (Figure 2.4).

140
130
120
Normalised activation (%)

110
100
90 Roy et al.
80 Tan et al.
70 Marras et al.
60
50
40
0 10 20 30 40 50 60
Lumbar range (degrees) Note: Standard deviation values were
not published for Tan or Marras

Figure 2.4 The effect of lumbar posture on the mean activation of LES

17
 As shown in Figure 2.4, the most flexed test position in the three studies examining
MIVC of ES was 50o. Based on average range of motion data from Pearcy & Hindle (1989),
who validated measuring lumbar range of motion with the Fastrak device against x-ray, it
is unlikely that subjects performed MIVC in full lumbar flexion. While other studies have
recorded ES muscle activation in end range lumbar flexion, their findings were not
published because activation levels were not the main focus of their study (Dolan &
Adams, 1993; Dolan et al., 1994). Therefore, it remains unclear as to how activation levels
of the paravertebral muscle may behave at the extreme of lumbar flexion during
maximum isometric contractions.

Previous studies have investigated EMG activation levels at the end range of
lumbar flexion using submaximal efforts and during dynamic back extension tasks (Dolan
& Adams, 1993; Dolan et al., 1994; Kaigle et al., 1998; Kippers & Parker, 1984; Marras,
Rangarajulu, & Wongsam, 1987; Toussaint et al., 1995). In these studies it has been found
that towards the end range of lumbar flexion ES muscle recruitment significantly
diminishes (Dolan et al., 1994; Kippers & Parker, 1984; Toussaint et al., 1995). Dolan et al.
(1994) found that when more than 80% of available lumbar motion had occurred, the LES
activation sharply declined. Others have shown a noticeable decline in activation of the
LES beyond 90% of flexion (Kippers & Parker, 1984). At the extreme of lumbar flexion it
has been frequently reported that the paravertebral muscles become electrically silent
even when carrying a load of 10-15kg (Dolan et al., 1994; Kippers & Parker, 1984;
Toussaint et al., 1995). Electrical silence of the paravertebral muscles when approaching
full flexion has come to be known as the ‘flexion-relaxation phenomenon’ (Dolan et al.,
1994). EMG fidelity is lost during dynamic movement (De Luca, 1997; Hof, 1984) and an
increase rate of change in lumbar curvature alters activation patterns of ES (Dolan et al.,
1994; Marras et al., 1984). Therefore, it remains unknown if the ‘flexion-relaxation
phenomenon’ would continue to exist at the extremes of flexion during maximal isometric
contractions.

While there is good evidence for lumbar posture influencing the level of muscle
activation of LES, less is known about the influence of lumbar posture on other divisions of

18
the paravertebral muscles such as the UES and multifidus. Dolan et al. (1994) recorded
activation of the UES as well as the LES during maximal isometric extensions from “slight”
lumbar flexion to the extreme of flexion. The results found the UES to be more activated
than the LES throughout all lumbar postures, although the difference was not statistically
significant. Toussaint et al. (1995) recorded activation of the UES, LES and also multifidus,
for six male participants while they performed a lowering and lifting tasks of a 15kg
weight. While the LES and multifidus displayed the ‘flexion-relaxation phenomenon’, the
UES remained consistently activated throughout the lifting cycle with no evidence of
relaxing at the extreme of flexion, irrespective of carrying a heavy load or not. It was
suggested that the different divisions of paravertebral muscles appear to function
independently of each other, indicating that recording activation at a single spinal level
may not be sufficient to represent all the different divisions of the paravertebral muscles
acting upon the lumbar spine. Whilst Toussaint et al. illustrated a different response
between the UES and LES with change in lumbar posture, no study to date appears to
have specifically reported on the influence of lumbar posture on levels of activation of the
different divisions of the paravertebral muscles when performing simulated lifting tasks at
maximal effort. Such information would assist in clarifying if the ‘flexion-relation
phenomenon’ continues to exist at end range of lumbar flexion during maximal lifting or
whether the spine remains under active muscular control.

2.4.4 Posture and neuromuscular efficiency

Expressing the amount of torque that can be produced per unit of muscle
activation (torque/activation), in any given lumbar posture can provide information on the
active (muscles) and passive (discs and ligaments) tissue contributions to extensor
moment (Tan et al., 1993). NME can be estimated from several studies where concurrent
measurements of maximal isometric torque and activation are reported (Marras et al.,
1984; Roy et al., 2003; Tan et al., 1993). In the study by Marras et al. (1984) maximal
isometric BE torque was measured in 0o, 22.5o and 45o of trunk inclination (the angle
between the trunk and thigh) concurrent with EMG activation of the lumbar ES. Data
were normalised against the maximum value found for each variable. Even though NME
was not directly calculated by the authors, using the available data, NME was found to rise
19
uniformly with an increase in flexion (Figure 2.5). NME was directly calculated by Tan and
co-workers (1993) using maximal BE torque and muscle activation of the medial ES
adjacent to L3, at 0o, 15o and 35o of lumbar flexion, with data normalised to the upright
posture (0o). As with Marras et al. a very similar NME trend was found (Figure 2.5). Roy
and colleagues (2003) also normalised maximal isometric BE torque and activation levels
of the LES to the upright posture to calculate NME between 0o and 50o of lumbar flexion.
There was an overall upward trend in NME with an increase in lumbar flexion, but after
10o of flexion the rate of change became less. Nonetheless, there appears to be a clear
pattern linking an increase in lumbar flexion with a rise in NME with end test point being
the most efficient (Figure 2.5).

1.8

1.6
NME (torque/activation)

1.4

1.2 Roy
Tan
1
Marras
0.8

0.6
0 10 20 30 40 50 60
Lumbar flexion (degrees)

Figure 2.5 Neuromuscular efficiency of LES

A number of theories have been developed to explain why neuromuscular

efficiency improves as lumbar spine flexion increases. While the relationship between
muscle activation and torque is considered to be linear (Dolan & Adams, 1993; Hof, 1984)
if torque production was proportional to activation in each posture, NME would remain
relatively unchanged. One theory giving rise to improved NME relates to muscle
elongation reducing the volume of muscle units below the electrode which diminishes the
amount of signal detected (De Luca, 1997; Tan et al., 1993). Other groups (Hof, 1984; Tan
et al., 1993; Ward et al., 2009) suggested that with an increase in lumbar flexion the
20
paravertebral muscles approach their optimal muscle length-tension, which is likely to
create more efficient torque production.

While torque improves with optimal muscle length-tension, torque is not solely
produced by active tissue alone, but also from tension within passive tissues (disc,
ligaments and connective tissues within muscles) (Bogduk, 2012; Dolan et al., 1994; Hill,
1938; Potvin et al., 1991). Dolan et al. (1994) investigated the passive tissue contribution
to BE torque throughout the lumbar flexion range. Measurements of maximal isometric
back extensions and activation of ES were recorded from a ‘slight’ flexed to fully flexed
lumbar posture. A series of equations was used to calculate the gradient of the
relationship between torque and activation in each posture and where the gradient
crossed the Y axis (no muscle activity), was considered the passive tissue contribution.
They concluded passive tissue contribution to be small from early to mid lumbar flexion
and then it becomes exponentially larger when more than 80% of available lumbar flexion
is exceeded. Passive tissue contribution at the extreme of lumbar flexion is also
supported by anatomical studies showing engagement of resistance of the ligaments and
disc towards the end of lumbar flexion (McGill, 1988).

With an exponential increase in passive tissue contribution towards the end of

lumbar flexion, a sharp increase in NME could be expected to be seen. In the studies
where NME has been estimated the highest degree of lumbar flexion used was 50o (Roy et
al., 2003). However, the NME at the extreme of flexion is unknown. At the end range of
lumbar flexion, high NME efficiency appears advantageous as more torque can be
produced with less muscle activation, leading to low metabolic demands. However, it is
the loading of the passive structures, in particular the interspinous ligament complex,
which increases anterior shear forces within the spine and this has the potentially to cause
tissue damage (McGill, 1988; Potvin et al., 1991).

The NME ratio of trunk extension has been calculated based on the LES muscle
activation with no known studies to date having determined the NME of the UES and
multifidus muscles in different lumbar postures. Dolan et al. (1994) have measured the
UES and LES activation levels at the extreme of flexion and have found the UES to remain
21
more activated than the LES at submaximal and maximal effort. The higher levels of
activation in the UES at the end of flexion would result in a lower NME than the LES, and
would indicate that the active contribution of the ES muscles to extensor torque in
extreme of lumbar flexion may be greater than previous thought. Considering the UES is
thought to contribute up to 74% of total BE torque (Bogduk, 2012), knowing its NME could
have a significant bearing of the overall NME of the lumbar spine.

2.5 Pelvis Fixation

Pelvis fixation is often used during measurements of BE torque and paravertebral
muscle activation to isolate trunk function from that of the pelvis (Dolan et al., 1994;
Graves et al., 1990a; Kaigle et al., 1998; Kippers & Parker, 1984; Parnianpour et al., 1991;
Roy et al., 2003; Smidt et al., 1983; Tan et al., 1993; Toussaint et al., 1995). Pelvis fixation
is proposed to isolate the trunk by preventing de-rotation of the pelvis and minimising the
contribution of the hip extensors (Graves et al., 1994; Walsworth, 2004). Without fixation,
it is thought the paravertebral muscles will only work isometrically to stabilise the spine
and it would be the hamstrings and gluteal muscles producing the prime movement of
extension (Graves et al., 1994; San Juan et al., 2005). However, the rationale for pelvis
fixation requires closer scrutiny and will be explored further in this section.

2.5.1 Pelvis fixation and back extensor torque

Considering the widespread use of pelvis fixation during BE torque studies (Dolan
et al., 1994; Roy et al., 2003; Smith et al., 2011; Tan et al., 1993), only one study was
identified testing the effect of pelvis fixation on BE torque (Smidt et al., 1983). Smidt et al.
(1983) tested 13 young participants and compared the effect of full pelvis fixation, partial
fixation (feet and thighs restrained but not the pelvis) and minimal fixation (only the feet
fixated) on BE torque at varying degrees of lumbar flexion, while seated on the Iowa trunk
dynamometer. While no significant difference was found between partial and full pelvis
fixation, a significant reduction in torque output was found when minimal fixation was
used.

22
 Pelvis fixation and torque production of the BE muscles has been examined in
studies investigating optimal BE muscle strengthening regimes (Da Silva, Lariviere,
Arsenault, Nadeau, & Plamondon, 2009; Graves et al., 1994; Smith et al., 2011; Udermann
et al., 1999). These studies tend to use isokinetic apparatus requiring participants to be
seated with and without pelvis fixation. Maximal BE torque was measured before and
after undertaking a strengthening regime using isokinetic machines where the level of
pelvis fixation was manipulated (Graves et al., 1994; Smith et al., 2011). Some participants
trained on the MedXTM whilst others were allocated to alternative machines considered to
give less pelvis fixation. Each group only trained once a week performing 8-12 repetitions.
The group that trained on the MedXTM significantly increased BE torque compared to
participants who trained on other devices. It was concluded the MedXTM was more
specific in targeting the paravertebral muscles due to pelvis fixation (Graves et al., 1994;
Smith et al., 2011). The infrequency of training undertaken made it difficult to determine
whether BE torque gains were due to pelvis fixation or if the MedXTM group improved due
to a training effect of using the MedXTM apparatus. Methodological issues with these
studies leave it unknown as to the effect pelvis fixation has on strengthening and torque
production of the BE muscles.

Udermann et al. (1999) also investigated the effect of pelvis restraints on BE

function and noted that it was virtually impossible to test lumbar extension torque in
sitting without some degree of pelvis fixation. This was also recognised in other studies
that compared full and partial fixation to no fixation (Da Silva et al., 2009; Walsworth,
2004). It is likely the apparatus used to measure BE torque dictated the need to fixate the
pelvis for its effective use, rather than fixation specifically targeting the paravertebral
muscles.

2.5.2 Pelvis fixation and muscle activation

A direct method for investigating the influence of pelvis fixation on BE muscles is
through muscle recruitment and amplitude measures using surface EMG. The effect of
pelvis fixation on recruitment of the paravertebral muscles has been examined in several
studies (Benson, Smith, & Bybee, 2002; Da Silva et al., 2009; San Juan et al., 2005;

23
Udermann et al., 1999; Walsworth, 2004), although only under concentric or eccentric
conditions, with no known studies specifically using isometric protocols. Within these
studies using dynamic protocols the results are conflicting on the effect of pelvis fixation
on ES activation. Two studies examined the effect of pelvis fixation on ES muscle
activation during dynamic back extension tasks performed at 40-50% of MVC in a seated
position (Da Silva et al., 2009; San Juan et al., 2005). Fixation was found to significantly
increase muscle activation in the UES and LES. However, partial and full pelvis fixation
resulted in similar activation levels, which in turn, were greater than those produced with
no fixation (San Juan et al., 2005). Conversely, in another study, the MedXTM was used to
measure activation of paravertebral muscles adjacent to L1 and L5 during maximal
isometric and dynamic back extension tasks. ES activation levels were similar for pelvis
fixation and no fixation conditions (Udermann et al., 1999; Walsworth, 2004). Walsworth
(2004) found no difference in levels of the LES activation during seated back extensions on
the MedXTM (thought to give ‘rigorous’ pelvis fixation) compared to the Cybex (which only
fixates the lower limbs and not the pelvis). It remains unclear as to the affect pelvis
fixation has on paravertebral muscles, but it does appear that activation levels do not
reduce with pelvis fixation.

Another purpose of pelvis fixation is to minimise the contribution of the hip

extensors in order to facilitate the recruitment of the BE muscles. Two studies used
surface EMG over the hip extensors to investigate the effect of pelvis fixation on hip
extensor recruitment during dynamic back extension tasks, using 40-50% of maximum
effort (Da Silva et al., 2009; San Juan et al., 2005). Results showed no significant
differences in activation levels of the hip extensors between pelvis fixation, partial fixation
or when the pelvis was free to move. Udermann et al. (1999) also used surface EMG on
the gluteal and hamstring muscle groups during dynamic back extensions using a load of
80% of body weight. They found no effect of fixation on hamstring muscle activation, but
the gluteals were found to be more active with the pelvis restrained compared to no
fixation. It remains unclear whether pelvis fixation can minimise the contribution of the
hip extensors during back extension tasks in order to facilitate the recruitment of the
paravertebral muscles.
24
2.6 Summary
From the literature review there are apparent gaps within the research pertaining
to the effect lumbar posture and pelvis fixation have on BE torque and paravertebral
muscle activation. The majority of studies investigating maximal BE torque and the effect
of lumbar posture have used a maximum of 50o of lumbar flexion. Considering that the
lumbar spine can have up to 70o of flexion in young healthy adults, less is known about BE
torque at the extremes of lumbar flexion. Furthermore, the vast majority of research
investigating BE torque is based upon experiments where participants are fixated in
isokinetic machinery and not while performing functional tasks such as lifting, leaving it
unknown if what we presently know about BE torque is transferable to functional tasks.

Muscle activation of ES has been shown to increase with lumbar flexion although
Marras et al. (1984) found contrasting results where ES activation reduced with flexion.
Akin to the research on extensor torque, maximum activation beyond 50 o of flexion
appears to be less understood. The majority of research has focused on ES activation at a
single spinal level, targeting LES. There is some evidence to suggest that the UES may
respond differently to lumbar postural changes compared to the LES but there is no
consensus within the literature, with little consideration given to the contribution of
multifidus.

Pelvis fixation is frequently used when examining the effect of lumbar posture on
BE torque and muscle activation in an attempt to isolate trunk function from that of the
pelvis and lower limbs. However, the ability of pelvis fixation to reduce the hip extensor
contribution to BE torque and/or to increase activation levels of the paravertebral muscles
appears to be conflicting between studies. Furthermore, fixating the pelvis is not
replicating functional positions or the actions of how the back and its muscle systems
would normally work together.

The disparities within the literature, as outlined above, create the purposes of this
study with the aim to extend present knowledge on the biomechanics of the lumbar spine.

25
2.7 Aims & Objectives
The aim of this study was to determine how lumbar posture and pelvis fixation
affected BE torque and level of muscle activation of three divisions of the paravertebral
muscles.

The objective of this study was to measure isometric BE torque and muscle
activation of three divisions of the paravertebral muscle (UES, LES and multifidus) in three
lumbar postures (fully lordotic, fully flexed and mid way between the former postures),
with and without the pelvis fixated, using a 3D force gauge and 3D motion analysis system
to measure BE torque, electromagnetic sensors to assess lumbar posture and surface EMG
to record muscle activation, respectively.

2.8 Purpose of the Study

The purposes of this study were based on its aims and objectives in order to
investigate the following hypotheses:

1. The effects of lumbar posture on BE torque. It was hypothesised that back

extensor torque would increase with lumbar flexion.
2. The effects of lumbar posture on paravertebral muscle activation. It was
hypothesised that muscle activation would initially increase with torque
production but reduce towards the end range of flexion.
3. The effects of lumbar posture on the UES, LES and multifidus muscle activity. It
was hypothesised that the relative muscle activity of each division of the
paravertebral muscles (UES, LES and multifidus) would increase with an
increase in lumbar flexion
4. The effect of neuromuscular efficiency (NME) of the paravertebral muscles. It
was hypothesised that NME would increase from the lordotic posture, to the
fully flexed posture.
5. The effect of pelvis fixation on isometric BE torque and paravertebral muscle
activation. It was hypothesised that pelvis fixation would increase torque
production and muscle activation.

26
 Chapter 3: Methods

3.1 Introduction
This chapter describes the methods used to investigate the hypotheses of this
study. The methods section will begin by outlining the study design and describing the
recruitment process of participants, which includes the criteria used for participant
selection. Following this, independent variables will then be defined and a detailed
description of kinetic, kinematic motion analysis and electromyography (EMG) measures
will be presented. The experimental procedure will be described along with data
processing methods and statistical analyses.

3.2 Methods
3.2.1 Study design
A repeated measures experimental design was used to investigate the effect of
lumbar posture (flexed, mid and lordotic) and pelvis fixation (fixed and free) on the
magnitude of BE torque and the level of three paravertebral muscle activation (adjacent
to L4/5, L3 and T10) during a maximal isometric trunk extension in a simulated lifting
position. A total of six conditions were performed. Ethical approval was gained from the
Auckland University of Technology Ethics Committee (AUTEC) (Appendix A).

3.2.2 Design overview

In this study participants were required to exert a maximal isometric back
extension, lasting three to five seconds, in three different lumbar postures. Postures
assumed were full lumbar flexion, maximal lumbar lordosis and the midpoint between
these extremes. The lumbar posture was determined using a Polhemus Fastrak device
(Polhemus Navigation, Kaiser Aerospace Inc., Vermont USA) with sensors attached to the
skin overlying L1 and S1 spinous processes. The Polhemus Fastrak device has been shown
to be a highly reliable measure of lumbar flexion and has been validated against x-ray
measurements of the lumbar spine (Adams & Dolan, 1991). Each posture was tested with
the pelvis free to move and with the pelvis restrained. A harness worn around the pelvis

27
with four straps anchored to the floor was used to fixate the pelvis (Montgomery,
Boocock, & Hing, 2011).

Estimates of BE torque were collected using two independent methods. The first
method utilized a three dimensional floor mounted force gauge that connected via a chain
to a chest harness worn by the participant. Maximal isometric back extension was exerted
through the chain to the force gauge. The second method, which occurred concurrently
with the first method, required the participant to stand with each foot on a separate force
plate while exerting the maximal back extension. The force plates measured ground
reaction forces for each lower limb independently.

Kinematic analysis was used to determine the position of the trunk, pelvis and
lower limbs during the maximal back extension. Thirty six retro-reflective markers were
placed over the body and equipment and their position in 3D space were captured by a
nine-camera motion analysis system (Qualysis Medical AB, Gothenburg, Sweden).

The kinetic and kinematic data were combined to estimate BE moment acting
about the lower back. The two methods used to calculate BE torque involved 1) force
measurements from the force gauge, hereafter referred to as the external moment, and
2) force measurements from two force plates, forming an eight segment rigid-link
biomechanical model, hereafter referred to as the internal moment.

Surface electromyography (EMG) of the UES, LES and multifidus muscles on the
right side of the spine were measured and synchronized with the kinetic and kinematic
data for the six different maximal isometric back extension test conditions. The level of
back muscle activation recorded during back extensions was normalized to maximum
isometric voluntary contraction performed in the Biering-Sorensen position.

3.2.3 Participants
Twenty-six healthy volunteers (13 males and 13 females) participated and
completed this study. The inclusion and exclusion criteria for the study were:

28
 Inclusion criteria:

 Aged 18-35 years old.

 Spoke and understood English.
 Self-reported to be medically fit and well.
 Able to attend one, two hour laboratory session at the Health
Rehabilitation Research Institute (HRRI) at AUT University.

Exclusion criteria:

 Previous spinal surgery.

 Previous abdominal surgery in the last 12 months.
 Medically diagnosed with any neurological or rheumatological condition.
 History of low back pain (LBP) in the previous 12 months that required
absence from work, LBP requiring medical treatment, or LBP lasted longer
than three days.
 Pregnant or given birth in the last 12 months.

The recruitment process is outlined in Figure 3.1. Participants were recruited via
word of mouth, poster advertisements on student and staff notice boards and
announcements on electronic notice boards at the university’s North Shore campus
(Appendix B). Individuals who expressed interest in participating in the study were
contacted by phone or by email to outline the requirements of the study and the risks
involved. Eligible volunteers were sent a copy of the participant information sheet to read
prior to testing (Appendix C). All participants who met the requirements of the study
were invited to attend the HRRI where the test procedure was explained in more depth
and a demonstration given, followed by an opportunity to ask questions. If the
participants were satisfied with the test procedure, a consent form was completed
(Appendix D) and an individual identification number allocated to each participant to
conceal identity.

29
 Participant information sheet given
n = 32
Excluded n = 3
(Scheurmanns, back
pain and knee injury)

Completed consent form and participated

n= 29
Excluded due to
technical difficulties
n= 3
Included in data analysis
n =26

Figure 3.1 Recruitment process

3.2.4 Randomisation process

Each participant was randomly assigned to the initial level of fixation (fixed versus
free) and the order in which each posture would be tested. There were six possible
posture sequence combinations available (Table 3.1). Microsoft Excel software package
was used to generate the random number sequences assigned to each pelvis fixation
condition and posture sequence.

Table 3.1 Pelvis fixation and posture sequence

Pelvis fixation Number Posture sequence Number

1 2 3
Fixed 1 Flexed, mid, lordotic 1

Free 2 Flexed, lordotic, mid 2

Mid, flexed, lordotic 3

Mid, lordotic, flexed 4

Lordotic, mid, flexed 5

Lordotic, flexed, mid 6

30
3.3 Experimental Measures
This study involved collection of kinetic data, kinematic data, measurement of
lumbar posture and recording EMG measures of the BE muscles. The following section
explains how these were undertaken.

3.3.1 Kinetic data

Two methods were used to measure force exertions:

A 3D force gauge (Advanced Mechanical Technology Inc., MC3A-100, USA) was

used to measure maximal isometric force exerted by the trunk when extending the back.
The force gauge was securely bolted to the floor located directly in front of the
participant’s feet. A custom-made metal plate with a loop attachment was bolted to the
top of the force gauge to allow a chain to be attached. A chain was attached between the
loop on the force gauge and the chest harness worn by the participant. A carabineer was
attached to the harness end of the chain so it could be detached in between trials to allow
the participant to return to an upright posture. The harness was made of woven non-
stretch material that had an adjustable Velcro TM chest belt and adjustable shoulder straps
to ensure the harness was located level with the sixth thoracic vertebrae (Figure 3.2). It
was worn as a close fit with micro fleece padding for participant comfort. The force gauge
data was collected at a sampling rate of 1200Hz and low pass filtered using a Butterworth
second order filter with a cut off frequency of 6Hz.

Figure 3.2 Chest harness marker position

31
 Ground reaction forces when performing back extensions were also measured
using two floor mounted force plates (Advanced Medical Technology Inc., USA). These
recorded 3D ground reaction forces independently for the right and left lower limbs of
each participant during maximal back extension exertions. These data were collected at a
sampling rate of 1200Hz, smoothed and passed through a Butterworth second order filter
with a cut off frequency of 6Hz due to the static nature of the test procedure.

3.3.2 Kinematic data

3.3.2.1 Motion analysis

Kinematic data of the trunk, pelvis and lower limbs were captured using a nine-
camera ProReflex motion analysis system (Qualysis Medical, Sweden) using a collection
rate of 60Hz. The motion analysis system was used to detect 36 retro-reflective markers
attached to the participant. The cameras were positioned to allow best visual coverage of
the experimental set up. The x, y, and z co-ordinate data from each retro-reflective
marker were subsequently stored in Qualysis software prior to further analysis using
Visual 3D biomechanics software package (C-motion Inc, version 4.0, USA).

3.3.2.2 Retro-reflective marker placement

Twenty retro-reflective markers (19mm diameter) were attached by an

experienced physiotherapist to the anatomical landmarks on the participant. Five markers
were also attached to the equipment, chest harness and force gauge, to identify its
relative 3D position within the laboratory (Figure 3.2 and Table 3.2). The retro-reflective
markers were attached with double backed adhesive tape (3M St.Pauls, MN, USA).
Reinforcement strips of FixamullTM were applied to each marker to prevent dislodgement
during testing. An additional 11 tracking marker were also attached to the thigh, shank
and trunk to assist in the identification and biomechanical modelling of each body
segment. These were chosen based on recommendations by Cappozzo, Della Croce,
Leardini, & Chiari (2005). While errors can arise from the placement of retro-reflective
markers on the skin surface with body movement, passive marker systems are non-

32
invasive, do not restrict natural movement patterns and have been shown to have a high
degree of accuracy (Cappozzo et al., 2005).

Figure 3.3 Adapted from Visual 3D; view of marker locations a. Posterior view b.
Anterior view

33
 Table 3.2 Anatomical location of retro-reflective markers and body segments
formed

Retro-reflective markers Type Body Segment

formed

Medial joint line of first metatarsal phalangeal joint B

Lateral joint line of fifth metatarsal phalangeal joint B

Posterior heel, three centimeters from floor B Foot

Distal tip of medial malleolus B

Distal tip of lateral malleolus B

Mid spine of tibia T Leg

Lateral knee joint line B

Medial knee joint line T

Thigh
Mid lateral thigh T

Greater trochanter B

Posterior superior iliac spine B Pelvis

Apex of iliac crest B

Fifth lumbar spinous process T

Tip of acromion B Thorax

Sternal notch T

Posterior chest belt E

Anterior chest belt E

Anterior superior iliac spine T

Additional markers
Top of the force gauge E

First thoracic spinous process T

B = body marker T = tracker marker E= equipment marker

34
 3.3.2.3 Biomechanical modelling

Markers were initially identified, tracked in the Qualysis software package

(Qualysis Track Manager) and exported in C3D format into Visual 3D (C-Motion Inc, USA,
version 4) for further biomechanical analysis. Visual 3D was used to create an eight
segment rigid-link, dynamic biomechanical model of the lower limbs, pelvis and trunk,
each body segment being represented as a known geometric objects scaled to each
participant (Figure 3.3) (Hanavan, 1964). For each body segment, the origin was
positioned at the proximal end of each body segment, the only exception being the trunk
which used the caudal end of the segment. Anthropometric data collected from each
participant ensured the model was as accurate as possible. Measurements used in the
construction of the model included body weight, height, pelvis depth level with the
anterior superior iliac spine, chest depth below the manubrium and the diameter of each
proximal thigh. BW estimates of segment masses and CoM were based on Dempster’s
(1955) data. The C3D files were exported from Visual 3D in ASCII format and saved in a
MicrosoftTM Excel file for subsequent statistical analysis.

Figure 3.4 The eight segment, rigid-link biomechanical model (from Visual 3D)

35
 3.3.2.4 Axes orientation

The force gauge, force plates and laboratory axes were orientated to match the
global movement co-ordinates in Visual 3D. The global movement co-ordinate system
was defined by co-ordinates in the three orthogonal axes:

 The anterior to posterior axis (X) was towards the midline of the two force
plates
 The medial to lateral axis (Y) was towards the direction the participant was
facing
 The vertical axis (Z) was directly upwards

The localised joint co-ordinate system for the feet, knees, hips, pelvis and trunk in
Visual 3D were aligned to the recommendations of the International Society of
Biomechanics Joint Co-ordinate systems (Wu et al., 2002). The local coordinate axes were
located at the proximal end of each segment, with the exception of the trunk where it was
located at caudad end of the segment. The three orthogonal axes were defined by:

 The anterior to posterior axis (X), directed forward from the participant
 The positive vertical axis (Y), directed upwards
 The medial to lateral axis (Z), directed to the right of the participant

3.3.2.5 Extensor moment calculation

The parameters used to estimate the external (extensor) moment are shown in
Figure 3.5. The extensor moment (EMX) was calculated after the addition of a force vector
to the trunk segment. The mass of the head and the upper body was based on the data
presented such that it was positioned 0.374 along the trunk segment axis. Its relative
position was based on the work of Winter (2009). EMX was estimated using the formula
adapted from Dolan & Adams (1993) and can be expressed as:

EMX = X1Fz + X2Bz + X3Hz

36
  FZ is the force applied to the force gauge.
 BZ is the mass of the trunk (0.355xBW (Dempster (1955)).
 Hz is the mass of the upper limbs and head.
 X1 is the horizontal distance between the anterior chest belt and the centre
of rotation (CoR) of the lumbar spine. The CoR of the lumbar spine was
located 7.5 centimeters anteriorly from the tip of the spinous process of
the fifth lumbar vertebrae (Dolan & Adams, 1993).
 X2 and X3 are the horizontal distances from the CoM of the trunk and the
CoM of the head and arms to the CoR of the lumbar spine, respectively
(Figure 3.5).

X2
X3

EM

CoM
CoR head/arms

Hz CoM
Bz trunk

X1 Fz
Force gauge

Figure 3.5 Parameters used to calculate extensor moment

Centre of mass of the upper body was calculated to be 37.4% inferior along the
line from the first thoracic vertebra spinous process to the line dissecting the greater
trochanters of the hip (Figure 3.6) (Dempster 1955). The CoR of the trunk segment was
located 0.5 along the segment length and determined by the default setting in Visual 3D.

37
 First thoracic vertebra

37.4%

Trunk centre of mass

62.6%

Greater Trochanter

Figure 3.6 Upper body centre of mass position

The internally derived moments, using the force measurements from the force
plates, were estimated using inverse dynamics in the unrestrained pelvis condition only.
Ground reaction forces were used to estimate joint reactions forces and muscle moments
acting about the ankles, knees, hips, and pelvis. The net muscle moment acting about the
pelvis represented the internally derived BE moment.

3.3.2.6 Lumbar posture

Lumbar posture was measured using the Polhemus Fastrak system (Polhemus
Navigation, Kaiser Aerospace Inc., Vermont USA). Fastrak is a real time, six degree-of-
freedom tracking system that detects rotation of a sensor located within the
electromagnetic field of a receiver unit. In the standing position, Fastrak 3D movement
sensors were secured using doubled sided adhesive tape (3M St.Paul, MN, USA) to the
skin surface superficial to the spinous process of the first lumbar vertebra (L1) and the first
sacral vertebra (S1). The L1 and S1 vertebral spinous processes were referenced from
locating the intervertebral space between L4 and L5, level with the top of the iliac crest
and cross checked by palpating up from S2 level located level with the PSISs (Snider,
Snider, Degenhardt, Johnson, & Kribs, 2011). FixomullTM tape was applied over the base
of the sensors to minimise their movement.

38
 Changes in lumbar posture were sampled at a frequency of 30Hz with LabVIEWTM
(National Instruments, Austin, Texas, USA, 2009, version 9.0). Lumbar curvature during lift
initiation was the relative angle between the L1 sensor relative to the receiver and
between the S1 sensor and the receiver (Dolan & Adams, 1993). A graphical
representation of lumbar curvature was available in real time for the participant to view.

3.3.2.7 Lumbar fixation

A custom-built harness was used to fixate the pelvis which enabled participants to
stand and adopt a lifting posture (Montgomery et al., 2011). The harness was made of
woven non-stretch material and was worn around the participant’s pelvis below the level
of the posterior superior iliac spines. The device was adjustable with a Velcro TM waist belt
and with a rubber inlay to prevent slippage. Four adjustable straps extended off the waist
belt to attach to anchor points on the floor. To provide the anchor points at the desired
locations, a wooden board was bolted to the laboratory floor with hooks attached at each
corner (Figure3.7). Weights were placed near the hooks to prevent the board lifting. The
central section of the board was removed to expose the force plates.

Figure 3.7 Fixation device and set up

39
 3.3.3 Electromyography
Surface electrodes were used to collect muscle EMG signals from the UES, LES and
multifidus. The following section describes how the skin was prepared, the electrode
placement determined and EMG was collected and processed.

3.3.3.1 Skin preparation

Prior to electrode placement L4/5 intervertebral space and L3 and T10 spinous
processes were located and marked with a pen on the skin surface by an experienced
physiotherapist. The skin surface superficial to an area two to three centimeters to the
right of midline was shaved and rubbed with alcohol wipes. The skin over the bony
prominence of the second thoracic vertebral spinous process was also prepared and an
earth electrode applied.

3.3.3.2 Electrodes and placement

Nortrode 20 silver/silver chloride 20mm bi-polar self-adhesive electrodes

(Nortrode 20TM, Myotronics, Inc, WA, USA) were applied to the prepared skin overlying
the UES, LES and multifidus. Skin resistance was measured for each bi-polar electrode and
a level of resistance less than 10kΩ was considered acceptable (De Luca, 1997).

Electrode placements were established according to the SEMTIL guidelines

(Hermans & Freriks, 2012) and the anatomical landmarks were located by an experienced
physiotherapist. The multifidus electrode was placed level with L4/5 intervertebral space,
on a 30o oblique orientation fanning more lateral, distally. the UES and LES electrodes
were attached to the skin surface 2-3 cm lateral and parallel to the spinous process of T10
and L3 spinous processes, respectively.

3.3.3.3 EMG collection and processing

EMG data were collected using a Bortec AMT8-channel differential amplifier

(Bortec Biomedical ltd, Canada). Each of the bi-polar electrodes were connected to a pre-

40
amplifier that plugged into the portable receiver located on a belt around the participant’s
waist. Pre-amplified signals from all the muscles of interest were transmitted, via a cable,
from the receiver unit to a Bortec Biomedical amplifier with variable gains of 1-3
continuous, 10GOhm input impedance and common mode rejection ratio of 115dB (De
Luca, 1997). The amplifier gain was set at 500 and the hardware high and low pass filters
were set at 10Hz and 500Hz, respectively. It is widely accepted within the literature that
most of the power spectrum of the EMG signal lies within 5-500Hz (Soderberg & Knutson,
2000). EMG signals from the amplifier were converted to digital data via the analogue to
digital conversion board and collected in the Qualysis software at a sampling frequency of
1200Hz. The EMG data were synchronized with the kinetic and kinematic data in the
Qualysis software.

The EMG data files were converted into text files and exported in text format for
analysis using LabVIEWTM (National Instruments, Texas, USA, version 9.0, 2009) software
programme. A visual cursor routine was used to establish a one second epoc where the
peak maximum force from the force gauge occurred during the three MVIC for each
posture. Visual inspection and power spectrum analysis of the epoc were undertaken to
ensure that no movement artifacts were evident. From the three trials for each posture,
the file with the highest mean force produced over the one second period was selected
for further data analysis in LabVIEWTM. The raw EMG data were demeaned and bandpass
filtered between 10 and 500Hz using a Butterworth 4th order filter.

The root mean square (RMS) method was used to assess the level of activation of
each muscle during rest, MIVC of back extension tasks. Briefly this method initially
squares all the data points within the epoc of EMG data in order to gain positive values for
all data points, and the final RMS value is estimated by calculating the square root of the
mean of the squared values. In order to compare levels of muscle activation across
muscles and individuals, RMS for MVIC were normalised and expressed as a percentage of
MVIC obtained during the Biering-Sorensen test. The normalization method of the EMG
(EMGN) subtracts the RMS of the EMG signal during rest from that during the lift initiation

41
task. The resulting value is divided by the difference in RMS recorded during MVIC and
rest, then multiplied by 100.

Lifting RMS - resting RMS

EMGN = X 100
Biering-Sorensen RMS - resting RMS

3.4 Experimental Procedures

3.4.1 Equipment calibration
At the start of each testing session the cameras were calibrated according to the
manufacturer’s instructions. The average movement residue of the markers during
calibration was less than two millimeters. The wooden board was bolted in situ around
the force plates and the location of the force plates and the force gauge were defined so
that the force vectors and laboratory coordinates were synchronized within Qualysis
software. Both force plates and the force gauge were zeroed prior to data collection.

3.4.2 Participant training

At the start of the test session all participants were familiarized with the testing
procedure and given training on how to assume the different lift initiation postures
(lordotic, mid and flexed positions) (Figure 3.8) and how to generate smooth, sustained
forces during the lift initiation task.

Figure 3.8 Lift initiation postures a. lordotic b. mid c. flexed.

42
 Visual feedback from a laptop connected to the Polhemus Fastrak device was used
to educate participants on how to assume each of the lumbar postures. A measurement
was taken from Fastrak of the range of motion of the lumbar spine from upright standing
to maximal forward flexion, whilst the participants maintained full knee extension. This
was recorded manually.

Participants were asked to connect the chain to the chest harness and were guided
into 45o of knee flexion to simulate lifting a box located in front of their feet. Once in this
position, the investigator used a combination of verbal and visual instructions along with
feedback from Polhemus Fastrak device to show participants how to anteriorly and
posteriorly tilt their pelvis. Participants practised until the extremes of lumbar flexion and
lordosis did not vary more than two or three degrees (as displayed visually on the laptop).
The range of motion from full lordotic to fully flexed lumbar posture in the lift initiation
position was recorded manually. Once achieved, the participant returned to the upright
standing posture to rest and repeated the process again to ensure consistency of range of
motion.

When the participant was proficient in assuming the three lumbar positions
(maximum lumbar lordosis, maximum lumbar flexion and the mid point of these two
extremes), they were asked to use approximately 50% to 75% of their maximum effort to
exert an upward tension from the chest harness on the chain connected to the force
gauge, while maintaining the different lumbar postures. During these sub-maximal
exertions participants were instructed to increase force over a three second period and
then to maintain a steady and constant force for a further three seconds. Participants
were given visual feedback of the force trace via a computer screen, allowing correction or
affirmation of the technique used. This was repeated a minimum of five times and until
the technique of maintaining lumbar curvature while sustaining a constant smooth force
was achieved. While the purpose of this session was to train the participant, it also served
as a warm-up exercise for the BE muscles.

43
 3.4.3 Participant preparation
At the start of the test session, participants were asked to change into close fitting
shorts and vest. The age of the participant was recorded and the anthropometric data
measured.

Anthropometric data included:

 Chest depth at mid sternum

 Right and left thigh width at the greater trochanter
 Height
 Body weight

The participant was required to lie prone on a plinth and surface electrodes were
applied to the skin surface superficial to the UES, LES and multifidus muscles on the right
side of the trunk. The participant was then asked to return to standing and the chest and
pelvis harness were fitted. The 36 retro-reflective markers were also applied to the skin
surface on the anatomical landmarks of the trunk and lower limbs.

Participants were asked to stand towards the front of the force plates, one foot on
each plate, equidistance from midline, shoulder width apart, to simulate a lifting a box
located in front of their feet. When foot position was established a line was drawn on the
floor around the participant’s feet to permit consistent foot position for each trial. The
Fastrak sensors were then attached to the L1 and S1 spinous processes.

Next, the chain length between the chest harness and force gauge was
determined. The participant was required to flex their knees to 45 o (as determined by a
manual goniometer). From this position the participant was instructed to flex forward and
simulate lifting the box from its handles, which were located 35cm above the floor (Figure
3.9). With the knees remaining in 45o of flexion, the lumbar spine posture was set to the
mid-position (as determined by the Fastrak), the participant applied light tension to the
chain, and the length of the chain connecting the chest harness to the force gauge was
determined.

44
 Figure 3.9 Method to establish chain length

3.4.4 Testing session

The first data file to be collected was a 10 second resting EMG file with the
participant lying prone on a plinth. A kinematic data file of the participant standing in the
anatomical position with each foot on a force plates was captured to act as the reference
‘static’ file to apply to subsequent dynamic files in Qualysis motion analysis.

The randomised sequence of posture tasks predetermined the initial fixation

condition and lumbar posture the participant was to adopt. The participant was
positioned with their feet on the designated foot marks, established during the practice
session, and instructed to connect the chain and carabineer from the force gauge to the
chest harness. Participants were asked to assume the required lumbar posture with the
aid of the Fastrak lumbar curvature angle displayed on the laptop screen. The participant
was instructed to exert light tension on the chain while keeping their knees flexed to 45o.
Once the participant confirmed they were in the correct lumbar posture an active motion
file within Qualysis was captured during the maximal isometric trunk extension. During
the maximal exertions participants were instructed to increase force over a three second
45
period and maintain a steady and constant force for a further three seconds. Verbal
encouragement was given throughout the maximal isometric contraction. At the end of
the each trial, the participant unclipped the chain from the chest harness, disconnected
the pelvis harness from the floor anchors and returned to an erect standing position. A
minimum rest period of two minutes was given between trials to allow sufficient time for
recovery and to minimise the effect of fatigue. The participant performed three maximal
isometric trunk extensions for each of the three lumbar postures (lordotic, mid and
flexion) in both the fixed and free pelvic conditions.

Following maximal isometric trunk extensions in all the lifting postures,

participants performed three maximum isometric back extensions in the Biering-Sorensen
position (Figure 3.10). The participant was positioned prone with their legs and pelvis
securely strapped to the plinth and their ASIS’s in line with the top of the plinth, leaving
the trunk unsupported. The chain from the chest harness was attached to the floor
mounted force gauge. The force and EMG data acquired during the MVIC performed in
the Biering-Sorensen position were used to normalize the data collection during the active
simulated lift trials. This position was chosen for normalisation because it has been shown
to be a reliable and commonly used method to normalise low back EMG data (Coorevits,
Danneels, Cambier, Ramon, & Vanderstraeten, 2008). Furthermore, data from the Fastrak
device indicated that lumbar curvature in the Biering-Sorensen position was similar to that
in upright standing.

Figure 3.10 Biering-Sorensen position

46
3.5 Statistical Analysis
All data were analysed for statistical significance using IBM SPSS version 19.0 (SPSS
Inc. Chicago) software package. Data for each trial were initially inspected using box plots
to statistically identify any potential outliers and also checked for normality. A two factor
analysis of variance (ANOVA) with repeated measures was used to test the main effect of
the two independent variables, posture (3) and pelvis fixation (2) on the dependent
variables of BE moment and muscle. Subsequent, 3x2x3 repeated ANOVA was conducted
for each of the three independent muscle divisions. The statistical significance was set to
an alpha level p<0.05. Significant main effects and interactions were subjected to post
hoc Bonferroni tests.

47
 Chapter 4: Results

4.1 Introduction
This chapter presents the findings from this study. It is divided into six sections
starting with the participants’ information, followed by how the data were managed. The
next section describes the findings related to the effect of posture on torque values and
muscle activation. This is followed by reporting the influence of posture on the
neuromuscular efficiency of the paravertebral muscles. The chapter is concluded by
discussing the effect of pelvis fixation on torque and muscle activation.

4.2 Participant Information

4.2.1 Demographic data
Table 4.1 displays the demographic data of the participants included in this study.
While the mean age of males and females were closely matched, the mean weight and
height were significantly greater for males compared to females (P>0.01).

Table 4.1 Participant demographics

Gender Mean age(yrs) Mean weight (kg) Height (cm)

Male (n=13) 24.2 78.50 176.6
Female (n=13) 22.6 64.35 163.8
Total 23.4 71.43 170.2

4.2.2 Lumbar range of motion

The Fastrak device measured the range of motion (ROM) of the lumbar spine from
relaxed upright standing to maximum forward flexion with knees extended (standing
ROM). Measurements of lumbar ROM in the lifting posture were also recorded between
the maximal lumbar lordotic position (end ROM of extension) and the maximal lumbar
flexion position, which is referred to as the lifting ROM (Figure 4.1).

Lumbar angle measurements were similar during maximal flexion in standing and
in the lift initiation position. However, when participants attempted to assume the

48
maximal lordotic posture in the lift initiation position the lumbar spine was flexed to 22o
when compared to the upright standing posture.

70
Max flex
60

50
Range of motion (degrees)

40

30

20
max lordotic
10

0
Standing ROM Lifting ROM
-10

Figure 4.1 Mean range of motion of the lumbar spine in standing and during lift
initiation position

4.3 Data Management

4.3.1 Back extensor torque calculation and validation
Descriptive statistics indicated that the mean peak torque of the externally derived
moment (force gauge) was slightly lower for all conditions than the internally derived
moment (force plates). However, independent t-tests showed no significant difference in
peak torque between the two methods of measurements (Table 4.2). Therefore, all
subsequent analysis of data used in this results section are expressed as torque values
calculated using the externally derived method (force gauge) as this method can be used
for calculation of torque in both free and fixated lifting postures.

49
 Table 4.2 T-test results comparing the two methods of torque calculation

Peak torque (Nm) Peak torque (Nm)

Test condition (External derived) (Internal derived) Sig level
Mean (SD) Mean (SD)
Free Lord 159.67 (55.5) 178.23 (48.4) 0.286
Free Mid 170.77 (48.6) 185.03 (46.9) 0.205
Free Flex 191.91 (51.7) 204.16 (46.4) 0.372

4.3.2 Torque and gender

ANOVA revealed main effects for gender and posture (P = 0.001), but no significant
gender by posture interactions. Subsequently post hoc analysis revealed that in all test
conditions males produced the highest mean torque irrespective of posture. The means
and standard deviations of the mean peak torque values collected for each gender in each
posture and pelvis fixation condition are displayed in Table 4.3. Flexed postures produced
the largest torque, followed by mid postures, with the least torque being produced in
lordotic postures for both genders.

Table 4.3 Mean (SD) peak torque for gender in each test condition

Torque (Nm)
Condition Males Females All participants
Mean (SD) Mean (SD) Mean (SD)
Fix Lord 210.23 (25.7) 136.25 (22.0) 173.24 (44.4)
Fix Mid 214.32 (30.8) 152.18 (25.6) 183.25 (42.11)
Fix Flex 220.18 (31.8) 176.37 (43.2) 198.28 (43.4)
Free Lord 212.84 (40.6) 143.61 (25.4) 178.23 (48.4)
Free Mid 220.09 (33.8) 149.97 (27.9) 185.03 (46.4)
Free Flex 234.33 (29.8) 174.00 (40.3) 204.16 (46.4)

50
 To take into account body mass differences between genders, peak torque was
expressed in terms of torque (Nm) per unit of body weight (kg) (Figure 4.2). Data analysis
revealed there were no significant differences (sig= 0.348) between males and females
when torque was normalised to body weight.

Body weight/ torque ratio (kg/Nm) 4

3.5
3
2.5
2
Male
1.5
Female
1
0.5
0
Lordotic Mid Flex
Posture

Figure 4.2 Torque expressed per unit of body weight for each posture and gender

Torque was also normalised by expressing torque as a percentage of maximum

torque produced in the Biering-Sorensen position (Figure 4.3). Again, data analysis
revealed there to be no gender differences when the data were normalised in this way.
Descriptive analysis showed when torque was normalised using the Biering-Sorensen
method compared to expressing torque as a percentage of body weight, males and female
torque data were similar. Therefore, male and female normalised torque data, expressed
as a percentage of MVIC during the Biering-Sorensen position are, hereafter, pooled for all
subsequent analysis of main effects and interactions.

51
 140

120

Normalised torque (%)

100

80

60 Male
Female
40

20

0
Lordotic Mid Flex
Posture

Figure 4.3 Torque expressed as a percentage of MVIC during the Biering-Sorensen

for each lumbar posture and gender

4.3.3 EMG and gender

The RMS value of the EMG data collected in each of the lifting postures was
normalised to the EMG recorded during MVIC in the Biering-Sorensen position. Although
males tended to activate the UES and multifidus at a greater percentage of MVIC than
females, the difference in the level of muscle activation was not significant between
genders (Figure 4.4). ANOVA also showed there was no significant interaction effect
between gender and posture, or between gender and level of fixation. Therefore, all
normalised EMG data presented, hereafter, are analysed collectively for male and
females.

52
 100
90

Muscle activation (% of MIVC)

80
70
60
50
Male
40
Female
30
20
10
0
UES LES Mult
Muscle division

Figure 4.4 Normalised EMG by muscle division for each gender

4.4 Lumbar Posture

4.4.1 Posture and torque/body weight ratio
Figure 4.5 shows that torque/body weight (BW) ratio increased from lordotic to
mid lifting posture and also from mid to flexed lifting postures. When torque was
normalised as Nm/Kg of BW, posture had a significant main effect on torque (p=0.001).
Post hoc analysis of comparing the three postures revealed that significantly higher torque
production occurred in the fully flexed position when compared to mid-range posture
(p=0.013), which in turn had significantly greater torque levels than when participants
adopted a lordotic lifting posture (p=0.014) (Table 4.4).

53
 3.2
*
3

BW/torque ratio (kg/Nm)

2.8 *

2.6

2.4

2.2

2
Lordotic Mid Flexed
Posture
*significant difference (P=0.001)

Figure 4.5 Torque expressed as a ratio of BW for each posture

Table 4.4 Post hoc analysis of torque/BW between different postures

ALL 95% CI
Posture1 Posture2 Sig. level Lower bound Upper bound
Flex Mid 0.013* 0.051 0.511
Flex Lordotic 0.001* 0.182 0.635
Mid Lordotic 0.014* 0.022 0.232

4.4.2 Posture and normalised torque

Figure 4.6 shows torque expressed as a percentage of the torque in Biering-
Sorensen position. A significant main effect of posture on normalised torque was found
(P>0.0001). Post hoc analysis found a significant difference between the flexed and mid
postures (p=0.004), mid and lordotic postures (p=0.014) and flexed and lordotic postures
(p=0.0001). Participants in the lordotic lifting posture produced 93% of the torque
generated during the Biering-Sorensen test, with a significant increase to 97% in the mid
posture, with a further increase of 11% from mid to fully flexed positions.

54
 140

130
* *

Normalised torque (%)

120

110

100

90

80

70

60
Lordotic Mid Flexed
Posture
* significant difference p =0.001

Figure 4.6 Normalised torque and posture

4.4.3 Posture and normalised EMG

Figure 4.7 shows EMG expressed as a percentage of the EMG activity in the
Biering-Sorensen position. ANOVA showed a significant main effect for posture (p=0.001)
and post hoc analysis (Table 4.5) revealed that there was a significant difference between
lordotic and mid postures (p=0.001). A 13% drop in activation occurred between the
lordotic and mid postures. A significant difference was also found between mid and
flexed postures (p=0.001), with a 41% decline in activation from the mid to flexed lifting
posture. The difference in activation between lordotic and flexed postures was also
significant (p=0.001), with lordotic postures producing the greatest amount of normalised
activation (Figure 4.7).

55
 120 * *
100

Normalised EMG (%)

80

60

40

20

0
Lordotic Mid Flexed
Posture
*Significant difference

Figure 4.7 Normalised EMG for each lifting posture

Table 4.5 Post hoc analysis significance levels and CI of normalised EMG

Posture1 Posture2 Sig. level 95% CI Upper bound

Lower bound
Flex Mid 0.0001* -36.28 -17.42
Flex Lordotic 0.0001* -45.53 -29.83
Mid Lordotic 0.001* -17.68 -3.98

4.4.4 Posture and normalised EMG by muscle division

Figure 4.8 displays the EMG for the UES, LES and multifidus in each lifting posture,
expressed as a percentage of the EMG produced in the Biering-Sorensen position. All
divisions of muscles exhibited a similar amount of activation for each posture, thus there
was no statistical significance between activation levels of each muscle division. All
muscle divisions showed a significant difference in the amount of activation between
flexed and mid postures, mid and lordotic postures and flexed and lordotic postures. All
muscle divisions showed the greatest decline in activation between the mid to flexed
postures.

56
 120 *
* *
100
* * *
*

Normalised EMG (%)

80 * *

60

40

20

0
Lord Mid Flex Lord Mid Flex Lord Mid Flex
UES LES Multifidus
Posture
* significant difference p =0.001

Figure 4.8 Mean and SD for normalised EMG for posture and muscle division

4.5 Neuromuscular Efficiency

Neuromuscular efficiency ratio was calculated by dividing normalised torque by
average normalised EMG of all three divisions (TN/EMGN). Normalised values were
calculated as a percentage of that produced in the of Biering-Sorensen position. There was
a significant main effect of posture on neuromuscular efficiency (p>0.0001). Post hoc
analysis revealed a significant difference between lordotic and mid postures (p=0.014),
mid and flexed postures (p=0.013), and subsequently, a difference between flexed and
lordotic postures (p=0.0001) (Table 4.6). NME was shown to increase by 23% between the
lordotic and mid lifting postures and by 87% from the mid to fully flexed lifting postures
(Figure 4.9).

Table 4.6 Post hoc analysis of posture on neuromuscular efficiency

95% CI
Postures compared Sig. level Lower bound Upper bound
Flex - Mid 0.013 0.051 0.511
Flex - Lordotic 0.0001 0.182 0.635
Mid - Lordotic 0.014 0.022 0.232

57
 3.5 *
3

NME ratio (Tn/EMGn)

2.5

2 *
1.5

0.5

0
Lordotic Mid Flexed
Posture
* significant difference p <0.01

Figure 4.9 The mean neuromuscular efficiency in each posture

Normalised EMG found each muscle division responded in the same way to each
other with alterations in posture. Consequently, the NME of each muscle division was
also comparable to the overall NME resulting in no significant effect of muscle division on
NME (Figure 4.10).

3.5
Neuromuscular efficiency (Tn/EMGn)

2.5
UES
2 LES
Multifidus
1.5

1
Lordotic Mid Flex
Posture

Figure 4.10 NME of each division of muscle in each posture

58
4.6 Pelvis Fixation
4.6.1 Fixation and normalised torque in each posture
Figure 4.11 shows the effect of pelvis fixation on torque for each posture. Whilst
the free pelvis condition generated more torque in each posture, the difference was small
therefore pelvis fixation was shown to have no significant effect.

115
110
Normalised torque (%)

105
100
95 Fix

90 Free

85
80
Lord Mid Flex
Posture

Figure 4.11 The effect of pelvis fixation condition in each posture

4.6.2 Fixation and normalised EMG for each muscle division

Figure 4.12 displays the normalised EMG for each muscle division under the fixed
and free pelvis conditions. Descriptive statistics showed that all muscles in the free
conditions produced slightly less activation than when the pelvis was free to move.
ANOVA revealed that there were no significant differences (p=0.38) in activation levels
between any of the muscle divisions when the pelvis was fixated or free to move.

59
 100
90
80

Normalised EMG (%)

70
60
50
Fixed
40
Free
30
20
10
0
UES LES Multifidus
Division of muscle

Figure 4.12 Normalised EMG and division of muscle comparing fixation

60
 Chapter 5: Discussion

5.1 Introduction
The purpose of this study was to investigate the effect of lumbar posture and
pelvis fixation on BE torque and the level of muscle activation of three separate divisions
of the paravertebral muscles during a simulated isometric lift initiation task. This chapter
aims to discuss the research findings and to compare it to the existing literature on this
topic. It will first begin by discussing the findings related to lumbar posture and its
influence on back extension torque and muscle activation, followed by NME. Pelvis
fixation and its effect on torque and muscle activation will be discussed. This section will
conclude by outlining the clinical implications of the study and its limitations.

5.2 Lumbar Posture

5.2.1 Posture and torque
The results of this study showed that posture had a significant effect on torque
production, with the flexed lumbar posture significantly producing more torque than mid
or lordotic postures. These findings concur with existing research that has found BE
torque to be greater in the flexed posture compared to other lumbar postures (Dolan et
al., 1994; Parnianpour et al., 1991; Raschke & Chaffin, 1996; Roy et al., 2003; Tan et al.,
1993; Toussaint et al., 1995). However, in contrast to previous research the current study
found that the greatest change in torque production occurred when moving from the mid
to flexed posture. Previous studies have found rapid increases in torque from a neutral
upright standing position to 20-30o of lumbar flexion (Parnianpour et al., 1991; Roy et al.,
2003; Tan et al., 1993), beyond which the rate of change decreases.

A possible reason for the contrasting findings between this study and those in the
literature may stem from differences in the methodological approaches used. Previous
research protocols have often been undertaken using isokinetic devices to test maximal
isometric BE torque with participants standing with their legs and pelvis fixated. The
upright standing position equated to the neutral posture (0o of flexion). However, the

61
current study used a more functional lifting posture with only the pelvis fixated and the
lower limbs free to move. When participants were asked to maximally extend their
lumbar spine in the functional lifting position their lumbar spine remained at
approximately, 22o of flexion (approximately one third of the way into flexion). While no
known research has specifically investigated lumbar range of motion in simulated lifting
postures, it is known that when performing activities that require hip flexion, there is
associated posterior pelvic rotation, which subsequently causes lumbar flexion
(Yasukouchi & Isayama, 1995). This flattening of the lumbar spine has also been noted
during dynamic lifting tasks (Arjmand, Plamondon, Shirazi-Adl, Lariviare, & Parnianpour,
2011; Potvin et al., 1991). The amount of lumbar flexion during simulated lifting using the
maximal lordotic posture in the current study was comparable to a mid-flexion position
used in previous studies. This also highlights that care should be taken when interpreting
findings from studies that have investigated the relationship between torque and lumbar
posture where the lumbar spine is extended or in a position of less than 20o of flexion.
Such postures are rarely seen during lifting tasks, particularly when lifting objects from
below knee height.

Another possible reason why the current study found a significant increase in
torque towards the end range of lumbar flexion may also relate to how posture influences
the length-tension relationship of ES muscle fascicles. Delp and colleagues (2001)
undertook an in vitro study to calculate optimal sarcomere length of Iliocostalis lumborum
and longissimus thoracis. A laser diffraction technique was used to establish muscle
sarcomere lengths in ES muscle fascicles, which were then normalised to the previously
established optimal sarcomere length of human muscle (Delp et al., 2001). It was
predicted that optimal fascicle length for the ES muscle group ranged from 16-20%
beyond that of resting length in supine. It was concluded that the ES muscles have an
improved ability to generate muscle force in more elongated (flexed) positions than in a
neutral (supine) posture. Optimal length-tension relationships in more flexed positions
have also been found in the multifidus muscle (Ward et al., 2009). Ward and colleagues
(2009) determined sarcomere length and the length-tension relationships of multifidus
fascicles that were dissected from patients undergoing surgery in flexed and lordotic
62
postures. Single fibre samples were attached to an ultrasensitive force transducer and a
micromanipulator that stretched the muscle fibre in 250 m increments. They found,
unlike other muscles within the body, the length-tension curve for multifidus continued to
increase as the spine became more flexed, reaching greatest tension at approximately 40o
of lumbar flexion. This suggests that multifidus has the potential to generate considerably
more tension when using postures in higher degrees of lumbar flexion.

The length-tension relationship of ES has also been estimated through

biomechanical modelling of participants undertaking back extension tasks and these
studies agree with the anatomical studies on how lumbar flexion influences torque
production (Mannion & Dolan, 1996; Raschke & Chaffin, 1996). For example, Raschke &
Chaffin (1996) recorded the EMG of ES at L3/4 whilst measuring dynamic trunk extensor
torque at 0o, 22.5o and 45o of trunk flexion. A standing frame was used to fixate the legs
and pelvis of participants while performing a lifting task which involved pulling up on a
handle attached to a floor mounted force gauge. A detailed anatomical biomechanical
model of the lumbar spine was used to estimate the length-tension of the ES. It was
concluded that the optimal length-tension of ES occurred at 45o of lumbar flexion.
Unfortunately, no measurements were taken beyond 45o of lumbar flexion, so it is unclear
what changes occur in the length-tension relationships for ES towards the end ranges of
lumbar flexion.

To date, there appears to be no studies published on maximal isometric extensor

torque for the extremes of lumbar flexion. The current study indicated that full lumbar
flexion was the most optimal position for extensor torque production. In full lumbar
flexion, the role of the passive tissues and their contributions to resisting bending
moments of the lumbar spine also needs further consideration. Biomechanical models
would suggest that there is significant loading on passive tissues of the lumbar spine at
the end range of spinal flexion, generated primarily from the iliolumbar ligaments and
intervertebral discs. Dolan et al. (1994) conducted a series of in vivo and in vitro
experiments to determine the contribution of the passive tissues of the lumbar spine
when resisting bending moments throughout lumbar flexion. They found that beyond

63
approximately 80% of full lumbar flexion there was an exponential rise of passive tissue
resistance to the bending moment. Dolan et al. concluded that the recruitment of the
passive structures of the lumbar spine may provide the potential energy to contribute to
the extensor moment at the end range of lumbar flexion.

Even though the current study found a significant effect of lumbar posture on BE
torque, the overall increase in torque in the lift initiation position when moving from a
fully lordotic to a fully flexed position was only 18%. Other studies have found 68%
(Parnianpour et al., 1991), 77% (Roy et al., 2003) and 83% increase (Tan et al., 1993).
These studies normalised torque to that measured in upright standing, which produced
considerably less torque than that measured in the Biering-Sorensen position used in this
study, even though the lumbar curvature was the same as that of upright standing. For
example, Roy et al., (2003) found a mean extension torque in standing of 101 Nm for a
group of young males and females (age range of 21-39 years), while the mean extension
torque in the present study was 182Nm. The lower levels of torque measured in an
upright standing posture may be associated with the postural stability demands required
during upright standing. For example, McGill (1991) measured maximal isometric
extension torque in standing and in the Biering-Sorensen position and found standing
postures to be an inferior method of testing optimal back performance. It was proposed,
in more complex tasks, such as in standing, the trunk muscles are responsible for
maintaining equilibrium in all three axes of movement of the spine and not solely in
producing torque in the desired direction. Thus, muscle exertion may be assigned to
stabilise the spine resulting in a net loss in torque in the desired direction. This
demonstrates the complexity of the spine and there is need for caution when applying
findings from non-functional measurements of BE torque to functional lifting tasks.

5.2.1.1 Gender and torque

The effect of gender on torque was found to be in agreement with previous

studies, with males being capable of generating higher peak torques than females (Dolan
et al., 1994; Graves et al., 1990b; Mannion & Dolan, 1996; Roy et al., 2003). However, as

64
others have reported (Keller & Roy, 2002), when torque was normalised to BW, the
current study found that peak torque did not significantly differ between genders.
Consideration of body dimension in torque production is also supported by anatomical
studies. For example, Mannion et al. (1997) obtained muscle biopsies from the UES and
LES of healthy young adults and found a positive correlation between body size and
muscle fibre size in both the UES and LES. This suggests that larger individuals have higher
ES muscle mass and therefore, greater potential to generate more extensor torque than
participants of a smaller size.

The gender findings of the current study differed to that of Graves et al. (1990b)
who found when BE torque was normalised against BW, males produced significantly
more BE torque per kilogram of BW than females. Differences in findings between the
present study and Graves et al. may relate to strength characteristics of male subjects in
each study. The mean maximum torque for males reported by Graves et al. was twice
that produced by the male participants in the current study, whereas females in each of
these studies produced similar extensor torque. A possible explanation for Graves et al.
findings may be related to the physical activity levels. The level of physical activity and
training of the BE muscles has been shown to be positively correlated to muscle
performance and ability to produce torque (Graves et al., 1994). Unfortunately, neither
the current study nor that of Graves et al. (1990b) measured physical activity levels
amongst their participants.

The current study found no differences between gender when BE torque was
normalised to that produced during maximal isometric trunk extension in the Biering-
Sorensen position. This test position was selected because it is commonly used for
strength testing of the trunk musculature (Biering-Sorensen, 1984; Coorevits et al., 2008)
and in present study the lumbar curvature was similar to that in upright standing. The
Biering-Sorensen position has also been shown to demonstrate a high level of consistency
for measures of BE torque (McGill, 1991). The present study’s findings are similar to Roy
et al. (2003), who normalised extension torque against maximum exertion in upright
standing and found this method to eliminate gender differences.

65
 5.2.3 Posture and EMG
Lumbar posture was found to have a significant effect on paravertebral EMG, with
greatest levels of activation in the lordotic posture and the lowest levels of activation
occurring when adopting the mid and flexed postures. Increases in paravertebral muscle
activation in the lordotic posture and diminishing activation towards the mid posture
(which equates to 22o to 40o of lumbar flexion respectively) are similar to those reported
by Marras et al. (1984) who found a relatively linear decline in the LES muscle activation
during maximal isometric lumbar extension from upright standing (0 o) to 45o of flexion.
However, muscle activity findings in the current study were contrary to the majority of
literature, which have shown a linear increase in muscle activity during MVIC of the BE
muscles in lumbar postures ranging from 0o (upright standing) to 50o lumbar flexion
(Dolan et al., 1994; Roy et al., 2003; Tan et al., 1993). These contrasting findings, may in
part, be explained by the positioning and fixation methods used in other studies. Roy et
al. (2003) and Tan et al. (1993) used the B200 isostation apparatus where participants
were required to stand with their pelvis and lower limbs fixated in a rigid frame with knees
fully extended. However, in the current study participants were fixated using a belt
system attached to the pelvis with no restraints placed on the lower limbs and allowing
the knees to flex to 45o.

An alternative explanation for a reduction in paravertebral muscle activation levels

from early in the range of flexion to 40oof lumbar flexion may be related to the level of
trunk support. The present study gave no external support to the trunk in attempt to
simulate a functional lifting task. In contrast, preceding research secured participants
upper torso to an isokinetic device for torque measurements. The additional support and
fixation of the torso may have resulted in altering the activation of the paravertebral
muscle as the trunk becomes more flexed. To some extent, this could explain why the
functional lifting task used in the present study resulted in a different activation pattern to
the findings reported in other studies that attempted to isolate the function of the lumbar
spine from the rest of the biomechanical chain.

66
 Another factor that should be taken into consideration when interpreting the
effects of lumbar posture on BE torque and paravertebral muscle activation is lumbar
stability. McGill (1991) investigated the muscle activation of the BE and abdominal
muscles during maximal isometric exertion with participants pelvis and trunk fixated in a
Cybex dynamometer whilst standing. It was found that in some postures there were low
levels of torque but high levels of ES and abdominal muscle co-contraction. It was
concluded that in some postures stabilising the lumbar spine during maximal exertions
was more important than the generation of torque. It is plausible that the recruitment
pattern of the paravertebral muscles during back extensions from flexed postures is aimed
at stabilising the spine when using low levels of lumbar flexion. Biomechanical models
have predicted that during functional tasks high stability demands occur in neutral or the
mid-range of spinal postures (McGill, 1991; Panjabi, Abumi, Duranceau, & Oxland, 1989).

A neurophysiological rationale for a reduction in muscle activation may be related

to the change in muscle length and its effect on the EMG signal. As muscles elongate
there is a reduction in the cross section area of the muscle fibres (Brown & Gerling, 2012).
EMG amplitude is proportional to the number of detectable firing motor units below the
electrode and therefore when a muscle elongates there is less volume below the
electrode, thus a reduction in activation (De Luca, 1997). This is unlikely to be the main
explanation for the decline in muscle activity with flexion in the current study as other
similar studies have shown EMG to increase in more flexed lumbar postures where the
paravertebral muscles have been elongated (Roy et al., 2003; Tan et al., 1993).

Further explanation for the increase in torque and a decrease in paravertebral

muscle activation between the lordotic and mid-range of lumbar flexion could be the
generation of torque by BE muscles other than those recorded in this study (Marras et al.,
1984; Phillips, Mercer, & Bogduk, 2008). Other muscles that have the potential to exert
an extension moment on the lumbar spine include quadratus lumborum, the hip
extensors and latissimus dorsi (LD). Phillips et al. (2008) investigated fascicle attachment
and orientation of quadratus lumborum and predicted that its capacity to generate BE
torque was limited to less than 10% of that of the paravertebral muscles. Conversely, the

67
hip extensors have a large torque generating capacity to assist in posteriorly rotating the
pelvis when raising the trunk and load being lifted, but they do not act directly across the
lumbar spine (Bogduk, 2012). Further evidence to support the role of LD and its
contribution to BE torque when lifting comes from anatomical studies. These have
revealed that up to two thirds of the force generating capacity of LD resides within the
lumbar region (Bogduk, 2012; Potvin et al., 1991). LD is considered to have only a modest
force generating capacity due to low number of sarcomeres arranged in parallel, but has
high numbers of sarcomeres arranged in series, allowing it to accommodate large changes
in length (Gerling & Brown, 2013). This sarcomere arrangement resulted in near maximal
force production across a wide range of lumbar motion including the end range of lumbar
flexion (Gerling & Brown, 2013; Raschke & Chaffin, 1996).

Several studies investigating the effect of posture on BE torque have investigated

the activation of LD (Marras et al., 1984; Raschke & Chaffin, 1996; Tan et al., 1993).
Marras et al. (1984) found that as the paravertebral muscle activation declined with
flexion, there was a concurrent increase in LD muscle activation. It was proposed that the
decline in activation of ES with lumbar flexion was compensated for by an increase in LD
recruitment. Furthermore, LD has been shown to be recruited more commonly when high
levels of trunk exertion are required (Tan et al., 1993) and particularly when performing
isometric contractions (Marras et al., 1984).

This study found a significant reduction in paravertebral muscle activation with

lumbar flexion. However, there was only a 14% reduction between the lordotic and mid
postures, compared to a 41% reduction in activation from mid to the full flexion. The
large reduction in ES muscle activation and increase in extensor torque between the mid
and the fully flexed posture suggest that towards end range of lumbar flexion torque
production is not generated solely by active tissue (muscle), but also through tension
within the passive spinal structures of the lumbar spine. Studies using biomechanical
modelling, which have assumed a linear relationship between torque and activation, have
predicted that a major contribution to BE torque during lumbar flexion comes from the
passive tissues of the spine (Dolan et al., 1994; Kippers & Parker, 1984). Findings from

68
these studies indicate that in the final 20% of lumbar flexion there is an exponential
increase in loading of the passive tissues (Dolan et al., 1994). It is possible that in the fully
flexed lumbar spine the forces arising from the passive tissue may result in inhibition of
muscle activation (Indahl, Kaigle, Reikeräs, & Holm, 1997), allowing a more metabolic
efficient and passive method of generating extensor torque.

The loading of passive tissues provides a likely explanation for the reduction in
activation between mid and fully flexed postures, but there is a lack of literature that has
investigated BE muscle activity during MVIC at the extremes of lumbar flexion. Studies
have been conducted where ES muscle activation has been recorded at end range of
lumbar flexion (Dolan & Adams, 1993; Dolan et al., 1994; Toussaint et al., 1995).
However, these studies have used submaximal isometric exertions or had participants
perform dynamic lifting tasks at submaximal effort. These studies showed ES muscle
activation to decline sharply in the last 20% of lumbar flexion, to such an extent that there
was total electrical silence of muscles at terminal flexion (Dolan et al., 1994; Kippers &
Parker, 1984; Toussaint et al., 1995). The present study also found a significant decrease
in paravertebral muscle activation at the end range of flexion but with recruitment
remaining at 39% of their maximum. These findings are similar to those of submaximal
lifting studies that have required participants to lift moderate weights between 10 and
15kg. (Dolan et al., 1994; Kippers & Parker, 1984; Toussaint et al., 1995). Toussaint et al.
(1995) noted ES muscle activation levels of up to 20% of maximum during the dynamic
lifting of a 15kg weight towards end range flexion.

5.2.4 Posture and EMG for each muscle division

All paravertebral muscle divisions demonstrated a similar reduction in activation
between lordotic to mid postures and from mid to flexed postures, differing by a
maximum of 5%. This was an interesting finding given that anatomical evidence would
suggest that the UES, LES and multifidus muscle architecture differs considerably, and it
has been implied that these muscle divisions may have varying roles when performing
functional tasks in different lumbar postures (Panjabi et al., 1989; Toussaint et al., 1995).
Whilst fibre orientation and location of the different division of ES may differ (Macintosh

69
et al., 1993), distribution of muscle fibre types in the UES and LES are similar (Mannion et
al., 1997). It has been argued that the similar fibre type distribution of the UES and LES
may reflect their similar function during a number of tasks.

Only one other study (Dolan et al., 1994) is known to have compared activation
levels of the UES and LES during a static lifting task. Dolan et al. (1994) recorded EMG of
the UES and LES during MVIC of BE muscles at 12 points of lumbar flexion while the
participant was pulling up on a floor mounted force gauge with their pelvis and lower
limbs restrained in a standing frame. Similar to the current study, no significant
differences were found between activation levels of the UES and LES in each lumbar
posture. Dynamic, submaximal lifting tasks have been used to investigated the effect of
posture on paravertebral muscle activation (Toussaint et al., 1995). It was found that
activation of the UES remained above 60% of MVC during the lifting task, with the greatest
activation occurring at full flexion. In contrast, muscle activation of the LES and multifidus
diminished in more flexed postures (Toussaint et al., 1995). It was concluded that towards
the end range of flexion the extensor torque produced by the LES is, in part, taken over by
the UES, mainly due to the greater mechanical advantage that the UES displays in this
position (Toussaint et al., 1995). However, care should be taken when interpreting results
from dynamic lifting tasks, because muscle activation and recruitment patterns are often
influenced by changes in angular velocity of the lumbar spine (Dolan et al., 1994; Hof,
1984).

The similar activation pattern of all three divisions, irrespective of their different
function, attachment and innervation indicates that during relatively static tasks the UES,
LES and multifidus are influenced by changes in posture in a similar way. This is also
supported by anatomical studies which have suggested that the overlapping arrangement
of the paravertebral muscle fascicles allow consistent forces to be generated as the spine
flexes (Macintosh et al., 1993). It is thought that a decrease or increase in mechanical
leverage of one fascicle is compensated by a reciprocal change in another (Macintosh et
al., 1993). This has implications for the rationale often used to justify the isolation of

70
individual ES muscles during exercise strengthening and functional rehabilitation
programmes (O'Sullivan, Twomey, & Allison, 1997).

5.2.5 Neuromuscular efficiency

This study found the neuromuscular efficiency (NME) ratio to increase from
lordotic to mid posture, with a greater increase occurring from mid to flexed postures.
Previous research also found a rise in NME with an increase in lumbar flexion. Both Roy et
al. (2003) and Tan et al. (1994) calculated the NME ratio of the LES in a neutral spine
posture to be 1.0, rising linearly to 1.6 at 50o of flexion and 1.67 at 35o of flexion. The
lordotic and mid postures in the present study represented approximately 20o and 40o of
lumbar flexion, with a NME ratio of 1.2 and 1.5 for the respective positions, which is
analogous to previous research. Unlike previous studies, which omitted to test the
extreme of lumbar flexion, this study found a much higher NME ratio of 2.8 in full flexion.
This represented an 87% increase in NME from the mid to end range of flexion, compared
to only 23% increase from lordotic to mid posture.

Tan et al. (1993) suggested improvements in the NME ratio of the LES, was in part,
the result of greater mechanical advantage of ES in mid-range flexion (35o) where
maximum torque was produced. However, Tan et al. only tested up to 35o and other
studies have suggested that 45o is the position of optimal length-tension of ES (Raschke &
Chaffin, 1996). While improvements in NME between the lordotic and mid posture could
be the result of improved mechanical advantage due to optimal length-tension of ES,
anatomical studies have demonstrated that at end range of lumbar flexion, the decrease
in ES muscle moment arm length has a reduced mechanical advantage by up to 39%
(Macintosh et al., 1993). These findings raises questions as to the mechanisms
contributing to a sharp increase in NME at end range of flexion, as found in this study.

The most likely reason for this study finding a large increase in NME ratio at the
end range of lumbar flexion can be explained by the change in the torque production from
the active to the passive tissues of the lumbar spine as it flexes. As discussed earlier in
this section, both biomechanical and anatomical studies have shown at end range of
flexion there is less torque produced from active tissues of the spine and there is greater
71
contribution to torque from the passive spinal structures, leading to relatively lower
metabolic cost (Bogduk, 2012; Dolan et al., 1994; Raschke & Chaffin, 1996; Tan et al.,
1993). While this study has shown end range of lumbar flexion to be the most
neuromuscular efficient, high torque production combined with low muscle activation can
potentially pose a high risk of back injury, suggesting high NME may not necessarily be
advantageous.

5.3 Fixation
5.3.1 Fixation and torque
This study found there was no effect on BE torque when the pelvis was restrained
compared to when the pelvis was free to move. Despite the wide spread use of pelvis
fixation in studies investigating BE torque and posture (Dolan et al., 1994; Graves et al.,
1990a; Kaigle et al., 1998; Kippers & Parker, 1984; Parnianpour et al., 1991; Roy et al.,
2003; Smidt et al., 1983; Tan et al., 1993; Toussaint et al., 1995), only one other study
appears to have investigated the efficacy of pelvis fixation on isometric BE torque. Smidt
et al. (1983) compared the effect of minimal fixation (only the feet restrained), partial
fixation (use of restraints across the thighs and lower leg but not the pelvis) and full
fixation on BE torque at varying degrees of lumbar flexion, while seated on the Iowa trunk
dynamometer. A significant reduction in extensor torque was found when using minimal
fixation compared to partial and full fixation. However, no difference in extensor torque
production was found between partial fixation and full fixation. These findings suggest
that providing some type of fixation to the thighs, and not specifically to the pelvis, is
important when using a trunk dynamometer. The present study did not fixate the lower
limbs, only the pelvis. Consequently, it is difficult to compare the results of this study with
those of Smidt et al., but the different forms of pelvis and lower body fixation may explain
why in this study no influence of pelvis fixation on BE torque was found.

There is a scarcity of research into the efficacy of using pelvis fixation, although
fixation has been widely used during studies investigating BE strength training regimes
(Graves et al., 1994; Smith et al., 2011). In these studies, it was noted that it was virtually
impossible to perform lumbar extension in sitting without some degree of fixation
72
(Udermann et al., 1999). This was also recognised in other studies comparing different
fixation configurations, where some degree of fixation was required as opposed to using
no restraints (Da Silva et al., 2009; Walsworth, 2004). It would appear that seated
isokinetic devices are more effective when using some degree of pelvis or lower limb
fixation and without fixation the ability to generate BE torque may be compromised.

The present study found no effect of pelvis fixation, which may be the result of
using a different and more functional position that does not require fixation in order to
perform the back extension task. Using isokinetic devices to measure BE torque appear to
require pelvis fixation and it is plausible that this method is chosen for the convenience of
testing and the experimental design. Unfortunately these devices are expensive, often
not accessible in the clinical setting and do not assess trunk extensor torque in functional
positions where the pelvis and lower limbs are free to move. In contrast, this study used a
functional task that could be performed equally as well with or without fixation. Testing
of the BE in functional positions seems important in order to study how the lumbar spine
works in conjunction with the pelvis, not just in isolation (Yasukouchi & Isayama, 1995).

5.3.2 Fixation and EMG

In this study, the level of pelvis fixation did not influence the level of paravertebral
muscle activation. No studies appear to have investigated the effect of pelvis fixation on
ES muscle activation during maximal isometric exertions. Research that has investigated
the effect of pelvis fixation on the activation of the BE muscles have often used dynamic
protocols. Findings from these dynamic studies are inconclusive, with some studies
finding increased levels of ES activation with fixation (Da Silva et al., 2009; San Juan et al.,
2005) while others have found no effect on activation levels of the paravertebral muscle
(Udermann et al., 1999; Walsworth, 2004).

The studies using dynamic protocols to investigate paravertebral activation, tend

to restrain the pelvis and lower limbs, which arguably gives rise to a higher degree of
stability than the method of fixation used in the present study; restraining only the pelvis.
McGill (1991) proposed that when spinal stability is challenged, an alteration in ES
recruitment pattern occurs in order to stabilise the spine in preference of generating
73
extensor torque. It was suggested that when more external stability is given in the form
of pelvis or lower limb fixation, this could potentially lead to alterations in ES activation
levels, although this was not found in the present study.

5.4 Clinical Implications

The findings of this study contribute to the body of knowledge involving
biomechanics of the lumbar spine. More specifically it has provided an understanding of
how lumbar posture influences BE torque and the activation of the paravertebral muscles
in a functional lifting position. This study has implications for people undertaking manual
handling lifting tasks, for organisations that provide manual handling guidelines as well as
health professionals involved in the management and prevention of LBI.

When a lordotic posture was maintained during a MVIC of the BE muscles, this
study found low levels of torque and high levels of paravertebral muscle activation. This
resulted in the lordotic posture having a low neuromuscular efficiency. In comparison, a
flexed lumbar posture was found to produce high levels of torque and low levels of
paravertebral muscle activation, showing that using a flexed lumbar spine when initiating
lifting has a high neuromuscular efficiency. The high levels of NME in more flexed
postures may provide some rationale for people choosing a flexed lumbar posture in
preference to a lordotic lumbar posture when performing lifting activities (Straker &
Duncan, 2000). The increased torque production in more flexed postures also has
implications for health professionals and highlights the need to standardize lumbar
posture when assessing BE torque.

The observed reduction in ES muscle activation concurrent to an increase in torque

between the lordotic and mid-range of flexion, suggests that the mid-range posture has
more optimal length-tension of the paravertebral muscles than a lordotic posture. In
more flexed positions the passive properties of ES and the other soft tissues of the spine,
as well as the recruitment of other muscles, may be important contributors to the
production of extensor torque. This has importance in the development of future
biomechanical models, which should consider the length-tension relationships and the

74
viscoelastic properties of other passive tissues of the spine when modelling BE torque and
ES activation.

Findings from the current study would indicate that using a mid-range posture
during maximal lifting has the benefit of generating higher BE torque than when using a
lordotic posture. The mid posture also has improved neuromuscular efficiency of the
paravertebral muscles compared to the lordotic spine, yet it does not excessively load the
passive spinal structures, which occurs at the end range of lumbar flexion. This study has
also shown that when lifting a box from below knee height, it appears that one third of
lumbar flexion is required, even when maintaining a maximally lordotic lumbar posture.
Advising individuals to lift without flexing the lower back appears unrealistic.

Previous research has shown paravertebral muscles have the ability to increase
their recruitment far into the flexion range when the participant is fixated to an isokinetic
apparatus. However, this study showed paravertebral muscle activation to reduce with
lumbar flexion, irrespective of pelvis fixation. The position of testing appears to be critical
in how the paravertebral muscles respond to changes in lumbar posture. This highlights
the importance of testing BE muscle strength in positions that simulate functional working
activities, in order to improve understanding on how the BE muscles would normally be
recruited. This also highlights the importance of using postures that simulate lifting
activities for functional training and rehabilitation of the paravertebral muscles of clients
involved in manual handling activities.

The findings of this study also have implications for the functional training of
individuals involved in manual handling activities. Traditional physiotherapy rehabilitation
techniques have focused on retraining individual divisions of ES and multifidus. The
functional postures used in this study appear to activate all divisions of the paravertebral
muscles equally and therefore, should be integrated into vocational and rehabilitation
programmes involving the lumbar spine.

Previous research investigating BE torque have used non-functional positions and

sought to isolate the lumbar spine by restraining the pelvis. The findings from this study

75
showed pelvis fixation to have no effect on BE torque or muscle activation, which has
implications for future studies into the biomechanics of lifting and questions the relevance
of using pelvis fixation. This also has implication for the clinical setting and questions the
justification for using fixation devices for functional strengthening and rehabilitating the
paravertebral muscles.

5.5 Limitations of the Study

This study was designed to be methodologically sound, although a number of
limitations have been identified.

Firstly, the study aimed to simulate maximal lifting ability in an initial lifting
posture. In order to estimate the extensor force acting on the torso and eliminate
variability in arm posture, a chain connected the chest harness to the force gauge. A more
natural lifting technique would have involved a lifting posture that used the upper limbs.

The data from three participants were excluded from the analysis due to retro-
reflective markers being hidden from the motion analysis cameras, or because markers
became dislodged. Marker displacement due to skin movement, particularly at the pelvis,
may have led to errors when determining body posture and therefore, should be
considered a potential source of error.

A main aim of this study was to investigate BE torque and the activation of the
paravertebral muscle group. However, the paravertebral muscles are not the only BEs,
with latissimus dorsi, gluteus maximus and quadratus lumborum also having the capacity
to generate an extensor moment on the lumbar spine. Determining the contribution of
these muscles to the BE torque was not feasible in the current study due to the
complexity of measuring EMG from these muscles. However, this was not seen as a major
limitation as the paravertebral muscles have been estimated to produce up to 80% of the
total extensor forces (Bogduk, 2012).

Pelvis fixation used in this study restricted pelvis movement through tension within
the harness and its anchor points on the floor. This allowed unrestricted movement of the

76
lower limbs and had the added advantage of enabling the retro-reflective markers to be
visible to the motion capture cameras. Although this proved effective in restraining the
pelvis, it was possible for the pelvis to move slightly during maximum back extension. A
study by Montgomery et al. (2011) using a similar pelvis restraint system and found
approximately 83% reduction of pelvic rotation during trunk flexion and rotation tasks.

Calculations of back extensor torque were calculated using two methods; internal
derived moment and external derived moment. These methods used a slightly different
centre of rotation for the lumbar spine: origin of the cylinder used to represent the trunk
segment (Internal); and a point approximating to the centre of the L5/S1 disc (external).
This may explain the slight systematic error in estimated torque resulting from the two
methods. Also, the centre of mass of the upper body was estimated using Dempster
(1955) data. It may have been preferable to estimate the centre of mass of the upper
body for each participant by directly measuring kinematic data using a marker set to
denote the upper body.

Finally, consideration should also be given to the fact that the task studied was
static and attempted to simulate an initial lifting posture. Whilst there is likely to be high
forces towards the initial stages of the lift, maximal effort may not occur during
movement. Therefore, extrapolating these findings to a functional lifting task should be
done with caution. It should also be borne in mind that this study was conducted on a
group of healthy young adults and therefore, these findings may not transpose to an older
population or those with a low back injury.

77
 Chapter 6: Conclusion & Recommendations

6.1 Conclusion
The main findings of this study showed that lumbar posture had a significant effect
on BE torque and paravertebral muscle activation. A posture simulating the initial stage of
a lift showed that when maintaining a lordotic posture during a MVIC of the BE muscles,
low levels of torque and high levels of paravertebral muscle activation occur.
Consequently, the lordotic posture had a low neuromuscular efficiency when compared to
mid-range and fully flexed lumbar postures. In contrast, performing the lift in a flexed
lumbar posture resulted in high levels of torque and low levels of paravertebral muscle
activation, showing that using a flexed lumbar spine has a high neuromuscular efficiency.
Neuromuscular efficiency at the end range of flexion was approximately twice that
reported in the literature, maybe because the majority of previous studies have not
investigated BE torque at the extremes of lumbar flexion.

While end range of lumbar flexion has high efficiency, the literature would suggest
that it is primarily achieved through the loading of passive spinal structures to resist
bending moments rather than additional muscle activation. The loading of the passive
tissues of the spine has been identified as a potential risk factor for LBI (Hoogendoorn et
al., 2000). Using a lumbar posture similar to the mid position used in this study is likely to
avoid the loading of the passive structures yet produce higher levels of neuromuscular
efficiency compared to the lordotic lumbar posture. Adopting a mid lumbar posture may
be useful in situations when exerting high BE torque is necessary.

During MVIC, the relative contribution of the UES, LES and multifidus were similar
in each lumbar posture tested. Traditional physiotherapy rehabilitation programmes have
often retrained a single division of the paravertebral muscles in non-functional positions
that do not reflect the lumbar postures adopted during lifting activities (O'Sullivan et al.,
1997). Findings from this study suggest that vocational and rehabilitation programmes
should use functional tasks that involve the combined movements of the trunk and lower
body in order to appropriately activate all divisions of the paravertebral muscles.
78
 Pelvis fixation was not found to influence BE torque or muscle activation. This
suggests that functional lifting tasks should not be concerned with restraining the pelvis
when measuring BE torque and paravertebral muscle activation. This study has
implications for the assessment of BE torque in the clinical setting, showing that pelvis
fixation may not be required when assessing trunk extension torque.

6.2 Recommendations
This study identified five key areas for future research.

i. Additional studies should be undertaken to establish the effects of lumbar

posture on torque and muscle recruitment patterns in an older population and
in people who have sustained a low back injury to direct rehabilitation
programmes more appropriately.
ii. Future studies should consider other BE muscles in addition to UES, LES and
multifidus, which contribute to BE torque in different lumbar postures, as this
may impact upon rehabilitation programmes.
iii. Further research should consider the response of BE muscles during the
performance of dynamic functional lifting tasks that normally occur in everyday
life.
iv. Additional research is required to establish the effect of asymmetrical trunk
postures on torque production and paravertebral muscle activation during
functional lifting tasks.
v. It would be of interest to health professionals for future research to establish
clinical based measures of BE torque in a clinical environment, without the
need for complex and expensive equipment.

79
 Reference List

ACC. (2010, 31.08.2010). ACC Injury Statistics 2008/2009. Section 14. Back Claims.
Retrieved 22.11.2010, 2010, from
http://www.acc.co.nz/PRD_EXT_CSMP/groups/external_ip/documents/reports_re
sults/wpc088576.pdf
Adams, M. A., & Dolan, P. (1991). A technique for quantifying the bending moment acting
on the lumbar spine in vivo. Journal of Biomechanics, 24(2), 117-126.
doi:10.1016/0021-9290(91)90356-R
Adams, M. A., Green, T. P., & Dolan, P. (1994). The strength in anterior bending of lumbar
intervertebral discs. Spine, 19(19), 2197-2203. doi:10.1097/00007632-199410000-
00014
Arjmand, N., Plamondon, A., Shirazi-Adl, A., Lariviare, C., & Parnianpour, M. (2011).
Predictive equations to estimate spinal loads in symmetric lifting tasks. Journal Of
Biomechanics, 44(1), 84-91. doi:10.1016/j.jbiomech.2010.08.028
Benson, M. E., Smith, D. R., & Bybee, R. F. (2002). The muscle activation of the erector
spinae during hyperextesion with and without the pelvis restrained. Physical
Therapy in Sport, 3(4), 165-174. doi:10.1054/ptsp.2002.0118
Biering-Sorensen, F. (1984). Physical measurements as risk indicators for low-back trouble
over a one-year period. Spine, 9(2), 106-119.
Bogduk, N. (2012). Clinical and Radiological Anatomy of the Lumbar Spine (5th edition
ed.). New York: Elsevier/Churchill Livingstone.
Bogduk, N., Macintosh, J. E., & Pearcy, M. J. (1992). A universal model of the lumbar back
muscles in the upright position. Spine, 17(8), 897-913. doi:10.1097/00007632-
199208000-00007
Bojadsen, T. W. A., Silva, E. S., Rodrigues, A. J., & Amadio, A. C. (2000). Comparative study
of Mm. Multifidi in lumbar and thoracic spine. Journal of Electromyography and
Kinesiology, 10(3), 143-149. doi:10.1016/S1050-6411(00)00004-3
Brown, S. H. M., & Gerling, M. E. (2012). Importance of sarcomere length when
determining muscle physiological cross-sectional area: A spine example.
Proceedings of the Institution of Mechanical Engineers, Part H: Journal of
Engineering in Medicine, 226(5), 384-388. doi:10.1177/0954411912441325
Callaghan, J. P., & McGill, S. M. (2001). Low back joint loading and kinematics during
standing and unsupported sitting. Ergonomics, 44(3), 280-294.
doi:10.1080/00140130118276
Cappozzo, A., Della Croce, U., Leardini, A., & Chiari, L. (2005). Human movement analysis
using stereophotogrammetry. Part 1: theoretical background. Gait & posture,
21(2), 186. doi:10.1016/j.gaitpost.2004.01.010
Cole, M. H., & Grimshaw, P. N. (2003). Low back pain and lifting: A review of epidemiology
and aetiology. Work, 21(2), 173-184.
Coorevits, P., Danneels, L., Cambier, D., Ramon, H., & Vanderstraeten, G. (2008).
Assessment of the validity of the Biering-Sørensen test for measuring back muscle
fatigue based on EMG median frequency characteristics of back and hip muscles.

80
 Journal of Electromyography and Kinesiology, 18(6), 997-1005.
doi:10.1016/j.jelekin.2007.10.012
Da Silva, R. A., Lariviere, C., Arsenault, A. B., Nadeau, S., & Plamondon, A. (2009). Pelvic
stabilization and semisitting position increase the specificity of back exercises.
Medicine and Science in Sports and Exercise, 41(2), 435-443.
doi:10.1249/MSS.0b013e318188446a
De Luca, C. J. (1997). The use of surface electromyography in biomechanics. Journal of
Applied Biomechanics, 13(2), 135-163. doi:10.1016/0021-9290(94)91124-X
Delp, S. L., Suryanarayanan, S., Murray, W. M., Uhlir, J., & Triolo, R. J. (2001). Architecture
of the rectus abdominis, quadratus lumborum, and erector spinae. Journal Of
Biomechanics, 34(3), 371-375. doi:10.1016/S0021-9290(00)00202-5
Dolan, P., & Adams, M. A. (1993). The relationship between EMG activity and extensor
moment generation in the erector spinae muscles during bending and lifting
activities. Journal of Biomechanics, 26(4-5), 513-522. doi:10.1016/0021-
9290(93)90013-5
Dolan, P., Mannion, A. F., & Adams, M. A. (1994). Passive tissues help the back muscles to
generate extensor moments during lifting. Journal of Biomechanics, 27(8), 1077-
1085. doi:10.1016/0021-9290(94)90224-0
Gallagher, S., & Marras, W. S. (2012). Tolerance of the lumbar spine to shear: A review and
recommended exposure limits. Clinical Biomechanics, 27(10), 973-978.
doi:10.1016/j.clinbiomech.2012.08.009
Gallagher, S., Marras, W. S., Litsky, A. S., & Burr, D. (2005). Torso flexion loads and the
fatigue failure of human lumbosacral motion segments. Spine, 30(20), 2265-2273.
doi:10.1097/01.brs.0000182086.33984.b3
Gerling, M. E., & Brown, S. H. M. (2013). Architectural analysis and predicted functional
capability of the human latissimus dorsi muscle. Journal Of Anatomy, 223(2), 112-
122. doi:10.1111/joa.12074
Graves, J. E., Pollock, M. L., Carpenter, D. M., Leggett, S. H., Jones, A., MacMillan, M., &
Fulton, M. (1990b). Quantitative assessment of full range-of-motion isometric
lumbar extension strength. Spine, 15(4), 289-294. doi:10.1097/00007632-
199004000-00008
Graves, J. E., Pollock, M. L., Foster, D., Leggett, S. H., Carpenter, D. M., Vuoso, R., & Jones,
A. (1990a). Effect of training frequency and specificity on isometric lumbar
extension strength. Spine, 15(6), 504-509.
Graves, J. E., Webb, D. C., Pollock, M. L., Matkozich, J., Leggett, S. H., Carpenter, D. M., ...
Cirulli, J. (1994). Pelvic stabilization during resistance training: Its effect on the
development of lumbar extension strength. Archives of Physical Medicine and
Rehabilitation, 75(2), 210-215.
Hanavan, E. P., Jr. (1964). A mathematical model of the human body. AMRL-TR. Aerospace
Medical Research Laboratories (6570Th), 1-149.
Hill, A. V. (1938). The heat of shortening and the dynamic constants of muscle. 136-195.
doi:10.1098/rspb.1938.0050
Hindle, R. J., Pearcy, M. J., Cross, A. T., & Miller, D. H. T. (1990). Three-dimensional
kinematics of the human back. Clinical Biomechanics, 5(4), 218-228.
doi:10.1016/0268-0033(90)90005-Q

81
Hof, A. L. (1984). EMG and muscle force: An introduction. Human Movement Science, 3(1-
2), 119-153. doi:10.1016/0167-9457(84)90008-3
Holmes, B., Leggett, S., Mooney, V., Nichols, J., Negri, S., & Hoeyberghs, A. (1996).
Comparison of female geriatric lumbar-extension strength: asymptotic versus
chronic low back pain patients and their response to active rehabilitation. Journal
of Spinal Disorders, 9(1), 17-22. doi:10.1097/00002517-199602000-00003
Hoogendoorn, W. E., Bongers, P. M., De Vet, H. C. W., Douwes, M., Koes, B. W., Miedema,
M. C., ... Bouter, L. M. (2000). Flexion and rotation of the trunk and lifting at work
are risk factors for low back pain: Results of a prospective cohort study. Spine,
25(23), 3087-3092. doi:10.1097/00007632-200012010-00018
Indahl, A., Kaigle, A. M., Reikeräs, O., & Holm, S. H. (1997). Interaction between the
porcine lumbar intervertebral disc, zygapophysial joints, and paraspinal muscles.
Spine, 22(24), 2834-2840. doi:10.1097/00007632-199712150-00006
Intolo, P., Milosavljevic, S., Baxter, D. G., Carman, A. B., Pal, P., & Munn, J. (2009). The
effect of age on lumbar range of motion: A systematic review. Manual Therapy,
14(6), 596-604. doi:10.1016/j.math.2009.08.006
Kaigle, A. M., Wessberg, P., & Hansson, T. H. (1998). Muscular and kinematic behavior of
the lumbar spine during flexion- extension. Journal of Spinal Disorders, 11(2), 163-
174. doi:10.1097/00002517-199804000-00013
Keller, T. S., & Roy, A. L. (2002). Posture-dependent isometric trunk extension and flexion
strength in normal male and female subjects. Journal of Spinal Disorders and
Techniques, 15(4), 312-318. doi:10.1097/00024720-200208000-00009
Kippers, V., & Parker, A. W. (1984). Posture related to myoelectric silence of erectores
spinae during trunk flexion. Spine, 9(7), 740-745.
Macintosh, J. E., & Bogduk, N. (1991). The attachments of the lumbar erector spinae.
Spine, 16(7), 783-792. doi:10.1097/00007632-199107000-00017
Macintosh, J. E., Bogduk, N., & Pearcy, M. J. (1993). The effects of flexion on the geometry
and actions of the lumbar erector spinae. Spine, 18(7), 884-893.
doi:10.1097/00007632-199306000-00013
Mannion, A. F., & Dolan, P. (1996). The effects of muscle length and force output on the
EMG power spectrum of the erector spinae. Journal of Electromyography and
Kinesiology, 6(3), 159-168. doi:10.1016/1050-6411(95)00028-3
Mannion, A. F., Dumas, G. A., Cooper, R. G., Espinosa, F. J., Faris, M. W., & Stevenson, J. M.
(1997). Muscle fibre size and type distribution in thoracic and lumbar regions of
erector spinae in healthy subjects without low back pain: normal values and sex
differences. Journal Of Anatomy, 190 ( Pt 4), 505-513. doi:10.1046/j.1469-
7580.1997.19040505.x
Marras, W. S., King, A. I., & Joynt, R. L. (1984). Measurements of loads on the lumbar spine
under isometric and isokinetic conditions. Spine, 9(2), 176-187.
doi:10.1097/00007632-198403000-00008
Marras, W. S., Lavender, S. A., Leurgans, S. E., Fathallah, F. A., Ferguson, S. A., Allread, W.
G., & Rajulu, S. L. (1995). Biomechanical risk factors for occupationally related low
back disorders. Ergonomics, 38(2), 377-410. doi:10.1080/00140139508925111
Marras, W. S., Rangarajulu, S. L., & Wongsam, P. E. (1987). Trunk force development
during static and dynamic lifts. Human Factors, 29(1), 19-29.

82
Mawston, G. A., & Boocock, M. G. (2012). Invited clinical commentary. The effect of
lumbar posture on spinal loading and the function of the erector spinae:
implications for exercise and vocational rehabilitation. New Zealand Journal of
Physiotherapy, 40(3), 135-140.
McBride, D., Begg, D., Herbison, P., & Buckingham, K. (2004). Low back pain in young New
Zealanders. The New Zealand Medical Journal, 117(1203), U1099-U1099.
McGill, S. M. (1988). Estimation of force and extensor moment contributions of the disc
and ligaments at L4-L5. Spine, 13(12), 1395-1402. doi:10.1097/00007632-
198812000-00011
McGill, S. M. (1991). Electromyographic activity of the abdominal and low back
musculature during the generation of isometric and dynamic axial trunk torque:
Implications for lumbar mechanics. Journal of Orthopaedic Research, 9(1), 91-103.
doi:10.1002/jor.1100090112
McGill, S. M., Hughson, R. L., & Parks, K. (2000). Changes in lumbar lordosis modify the
role of the extensor muscles. Clinical Biomechanics, 15(10), 777-780.
doi:10.1016/S0268-0033(00)00037-1
Montgomery, T., Boocock, M., & Hing, W. (2011). The effects of spinal posture and pelvic
fixation on trunk rotation range of motion. Clinical Biomechanics, Article in press.
doi:10.1016/j.clinbiomech.2011.02.010
Ng, J. K., Kippers, V., Parnianpour, M., & Richardson, C. A. (2002). EMG activity
normalization for trunk muscles in subjects with and without back pain. Medicine
& Science in Sports & Exercise, 34(7), 1082-1086.
O'Sullivan, P. B., Twomey, L. T., & Allison, G. T. (1997). Evaluation of specific stabilizing
exercise in the treatment of chronic low back pain with radiologic diagnosis of
spondylolysis of spondylolisthesis. Spine, 22(24), 2959-2967.
doi:10.1097/00007632-199712150-00020
Panjabi, M., Abumi, K., Duranceau, J., & Oxland, T. (1989). Spinal stability and
intersegmental muscle forces. A biomechanical model. Spine, 14(2), 194-200.
doi:10.1097/00007632-198902000-00008
Parnianpour, M., Campello, M., & Sheikhzadeh, A. (1991). The effect of posture on triaxial
strength in different directions: Its biomechanical consideration with respect to
incidence of low-back problem in construction industry. International Journal of
Industrial Ergonomics, 8(3), 279-287. doi:10.1016/0169-8141(91)90038-N
Parnianpour, M., Li, F., Nordin, M., & Kahanovitz, N. (1989). A database of isoinertial trunk
strength tests against three resistance levels in sagittal, frontal, and transverse
planes in normal male subjects. Spine, 14(4), 409-411.
Pearcy, M. J., & Hindle, R. J. (1989). New method for the non-invasive three-dimensional
measurement of human back movement. Clinical Biomechanics, 4(2), 73-79.
doi:10.1016/0268-0033(89)90042-9
Phillips, S., Mercer, S., & Bogduk, N. (2008). Anatomy and biomechanics of quadratus
lumborum. Proceedings of the Institution of Mechanical Engineers, Part H: Journal
of Engineering in Medicine, 222(2), 151-159. doi: 10.1243/09544119JEIM266
Potvin, J. R., McGill, S. M., & Norman, R. W. (1991). Trunk muscle and lumbar ligament
contributions to dynamic lifts with varying degrees of trunk flexion. Spine, 16(9),
1099-1107. doi:10.1097/00007632-199109000-00015

83
Punnett, L., Fine, L. J., Keyserling, W. M., Herrin, G. D., & Chaffin, D. B. (1991). Back
disorders and nonneutral trunk postures of automobile assembly worker.
Scandinavian Journal of Work, Environment and Health, 17(5), 337-346.
doi:10.5271/sjweh.1700
Raschke, U., & Chaffin, D. B. (1996). Support for a linear length-tension relation of the
torso extensor muscles: an investigation of the length and velocity EMG-force
relationships. Journal Of Biomechanics, 29(12), 1597-1604. doi:10.1016/S0021-
9290(96)80011-X
Roy, A. L., Keller, T. S., & Colloca, C. J. (2003). Posture-dependent trunk extensor EMG
activity during maximum isometrics exertions in normal male and female subjects.
Journal Of Electromyography And Kinesiology: Official Journal Of The International
Society Of Electrophysiological Kinesiology, 13(5), 469-476. doi:10.1016/S1050-
6411(03)00060-9
San Juan, J. G., Yaggie, J. A., Levy, S. S., Mooney, V., Udermann, B. E., & Mayer, J. M.
(2005). Effects of pelvic stabilization on lumbar muscle activity during dynamic
exercise. Journal of Strength and Conditioning Research, 19(4), 903-907.
doi:10.1519/R-15684.1
Singh, D. K. A., Bailey, M., & Lee, R. Y. W. (2011). Ageing modifies the fibre angle and
biomechanical function of the lumbar extensor muscles. Clinical Biomechanics,
26(6), 543-547. doi:10.1016/j.clinbiomech.2011.02.002
Smidt, G., Herring, T., Amundsen, L., Rogers, M., Russell, A., & Lehmann, T. (1983).
Assessment of abdominal and back extensor function. A quantitative approach and
results for chronic low-back patients. Spine, 8(2), 211-219. doi:10.1097/00007632-
198303000-00014
Smith, D., Bissell, G., Bruce-Low, S., & Wakefield, C. (2011). The effect of lumbar extension
training with and without pelvic stabilization on lumbar strength and low back
pain. Journal of Back & Musculoskeletal Rehabilitation, 24(4), 241-249.
doi:10.3233/BMR-2011-0301
Snider, K. T., Snider, E. J., Degenhardt, B. F., Johnson, J. C., & Kribs, J. W. (2011). Palpatory
Accuracy of Lumbar Spinous Processes Using Multiple Bony Landmarks. Journal of
Manipulative and Physiological Therapeutics, 34(5), 306-313.
doi:10.1016/j.jmpt.2011.04.006
Soderberg, G. L., & Knutson, L. M. (2000). A guide for use and interpretation of
kinesiologic electromyographic data. Physical Therapy, 80(5), 485-498.
Straker, L., & Duncan, P. (2000). Psychophysical and psychological comparison of squat
and stoop lifting by young females. Australian Journal of Physiotherapy, 46(1), 27-
32.
Tan, J. C., Parnianpour, M., Nordin, M., Hofer, H., & Willems, B. (1993). Isometric maximal
and submaximal trunk extension at different flexed positions in standing: triaxial
torque output and EMG. Spine, 18(16), 2480-2490. doi:10.1097/00007632-
199312000-00018
Toussaint, H. M., De Winter, A. F., De Haas, Y., De Looze, M. P., Van Dieen, J. H., & Kingma,
I. (1995). Flexion relaxation during lifting: Implications for torque production by
muscle activity and tissue strain at the lumbo-sacral joint. Journal of Biomechanics,
28(2), 199-210. doi:10.1016/0021-9290(94)00051-5

84
Udermann, B. E., Graves, J. E., Donelson, R. G., Ploutz-Snyder, L., Boucher, J. P., & Iriso, J.
H. (1999). Pelvic restraint effect on lumbar gluteal and hamstring muscle
electromyographic activation. Archives of Physical Medicine and Rehabilitation,
80(4), 428-431. doi:10.1016/S0003-9993(99)90280-0
Walsworth, M. (2004). Lumbar paraspinal electromyographic activity during trunk
extension exercises on two types of exercise machines. Electromyography and
Clinical Neurophysiology, 44(4), 201-207.
Ward, S. R., Kim, C. W., Eng, C. M., Gottschalk, I. V. L. J., Tomiya, A., Garfin, S. R., & Lieber,
R. L. (2009). Architectural Analysis and Intraoperative Measurements Demonstrate
the Unique Design of the Multifidus Muscle for Lumbar Spine Stability. Journal of
Bone & Joint Surgery, American Volume, 91-A(1), 176-185.
doi:10.2106/JBJS.G.01311
Waters, T. R., Putz-Anderson, V., & Garg, A. (1994). Applications manual for the revision of
NIOSH lifting equation. Cincinnati. Retrieved from
http://www.cdc.gov/niosh/docs/94-110/
Wu, G., Siegler, S., Allard, P., Kirtley, C., Leardini, A., Rosenbaum, D., ... Stokes, I. (2002).
ISB recommendation on definitions of joint coordinate system of various joints for
the reporting of human joint motion—part I: ankle, hip, and spine. Journal Of
Biomechanics, 35(4), 543-548. doi:10.1016/S0021-9290(01)00222-6
Yasukouchi, A., & Isayama, T. (1995). The relationships between lumbar curves, pelvic tilt
and joint mobilities in different sitting postures in young adult males. Applied
Human Science: Journal Of Physiological Anthropology, 14(1), 15-21.

85
 Appendices

Appendix A

MEMORANDUM
Auckland University of Technology Ethics Committee (AUTEC)

To: Grant Mawston

From: Dr Rosemary Godbold and Madeline Banda Executive Secretary, AUTEC
Date: 10 May 2011
Subject: Ethics Application Number 11/15 The effect of lumbar spine
position on back extensor torque and trunk muscle activation.

Dear Grant
Thank you for providing written evidence as requested. We are pleased to advise
that it satisfies the points raised by the Auckland University of Technology Ethics
Committee (AUTEC) at their meeting on 14 February 2011 and that on 19 April 2011, we
approved your ethics application. This delegated approval is made in accordance with
section 5.3.2.3 of AUTEC’s Applying for Ethics Approval: Guidelines and Procedures and is
subject to endorsement at AUTEC’s meeting on 23 May 2011.
Your ethics application is approved for a period of three years until 19 April 2014.
We advise that as part of the ethics approval process, you are required to submit
the following to AUTEC:
 A brief annual progress report using form EA2, which is available online through
http://www.aut.ac.nz/research/research-ethics/ethics. When necessary this form
may also be used to request an extension of the approval at least one month prior
to its expiry on 19 April 2014;
 A brief report on the status of the project using form EA3, which is available online
through http://www.aut.ac.nz/research/research-ethics/ethics. This report is to
be submitted either when the approval expires on 19 April 2014 or on completion
of the project, whichever comes sooner;
It is a condition of approval that AUTEC is notified of any adverse events or if the
research does not commence. AUTEC approval needs to be sought for any alteration to
the research, including any alteration of or addition to any documents that are provided
to participants. You are reminded that, as applicant, you are responsible for ensuring that
research undertaken under this approval occurs within the parameters outlined in the
approved application.
86
 Please note that AUTEC grants ethical approval only. If you require management
approval from an institution or organisation for your research, then you will need to make
the arrangements necessary to obtain this.
When communicating with us about this application, we ask that you use the
application number and study title to enable us to provide you with prompt service.
Should you have any further enquiries regarding this matter, you are welcome to contact
Charles Grinter, Ethics Coordinator, by email at ethics@aut.ac.nz or by telephone on 921
9999 at extension 8860.
On behalf of AUTEC and ourselves, we wish you success with your research and
look forward to reading about it in your reports.
Yours sincerely

Dr Rosemary Godbold and Madeline Banda

Executive Secretary
Auckland University of Technology Ethics Committee
Cc: Laura Holder laura.holder@aut.ac.nz, Peter McNair

87
 Appendix B

An investigation of back
strength and muscle activation
in different postures.
Volunteers required!
 This is part of a Physiotherapy research project
investigating back muscle strength and
activation in different postures.
 Participants must be aged between 18 and 35
years old.
 We welcome participants with or without a
history of back pain.
 You should not have any other muscle, joint, or
neurological disorder.
 You should not have received any spinal or
abdominal surgery in the past.
 You will not be included in the study if you are
pregnant.
 You are required to be able to speak and
understand English.

For further information please contact Laura on

laura.holder@aut.ac.nz or 021 488430.

88
Appendix C

Participant
Information Sheet
Date Information Sheet Produced:
05. 07. 2011
Project Title
The effect of lumbar spine position on back extensor torque and trunk muscle activation.
An Invitation
My name is Laura Holder and I am a physiotherapist. I would like to invite you to take part in this
study that is investigating the effect of lower back position on how different parts of the back
muscles are activated and produce force. I am undertaking this research as part of a Masters of
Health Science. Before you accept this invitation please read the following outline of the study.
You will need to make a decision about taking part in this study prior to attending the testing
session. Your participation is entirely voluntary and you can withdraw from the study at any time
without giving a reason.
What is the purpose of this research?
This purpose of this research is to determine if different parts of the back muscles work differently
when the back position is changed and also when the pelvis is restricted from moving compared to
when it is free. It will mainly be studied in people who have no back pain, with four or five case
studies on people who have low back pain. Since low back pain affects one in two New Zealanders
and most injuries are associated with a forward bent position, knowing how muscles work may
impact upon the rehabilitation that physiotherapist plan for their back pain clients. The successful
completion of this study may lead to the information being presented at professional conference or
within a professional journal. It will also give the principle researcher a Masters of Health Science
qualification.
How was I identified and why am I being invited to participate in this
research?
You as a participant are welcomed to take part in this research because you have responded to
advertisement requesting volunteers for this study. We are seeking volunteers who are aged 18-
35yrs old, whether or not you have back pain. You will be excluded if you have had previous spinal
or abdominal surgery, generalized muscle joint or neurological disorders, have a medical condition
that may preclude maximum effort strength testing, are pregnant or unable to speak and
understand English.
Participants must also ‘fall’ into one of two categories:
1. If you have not experience any pain originating from the back in the last 12 months.

89
or
2. If you have low back pain, your symptoms must have been present for at least three
months in the last 12 months, have experienced pain in the last month but not have radiating leg
symptoms.

What will happen in this research?

Your participation will involve attending the AUT University North Shore Campus for one session of
data collection. It is advised to come dressed in loose fitting clothes and wear shorts. You will be
first required to complete a consent form and there will be opportunity to ask me any questions
about what is involved. You will then be required remove your top (a modesty gown will be
provided so your front will remain covered but exposures the skin on your back). Next, you will be
asked to lie on a bed on your front and have three pads/electrodes attached to the right side of your
back muscles and one to the front of your tummy muscles, to measure their activity. You will then
be required to undertake a 10 minute warm-up consisting of five minutes on a static bike followed
by practice lifts at 50 percent of your maximum effort in the test position. Two small movement
sensors will then be placed on your lower back to measure spinal motion (see figure 1) and this
connects to the laptop. Approximately 30 small reflective markers will be placed on your legs, and
trunk that will enable a nine-camera system to detect your body position. This will allow me to
convert the reading on the force gauge in to torque measurement. Testing will then begin. With
your knees semi-bent you will be asked to stoop over as though you are about to lift a box off the
floor. The chest harness that I would like you to wear will have a small chain connecting it to the
force gauge of the floor. Maintaining this position your back muscles will be tested statically in three
different back positions; maximum bent, maximum arched and half way between. The laptop will
give you feedback so you will be able to maintain the correct position. You will be tested under two
different conditions. One is with the pelvis fixed (see figure 1a) and the other without pelvis fixation
(see figure 1b). In each of these positions you will be required to pull up on the chain and sustain a
maximum effort for five seconds. This will be repeated three times in each position with a two
minute rest where you will be able to stand upright. In total, you will be required to perform 18
maximum contractions lasting five seconds each.

1a. Fixed test position 1b. Unfixed test position

Figure 1: Diagram of the test position.
What are the discomforts and risks?
The risks involved are minimal. However, when doing any unfamiliar activity it can result in mild
muscle discomfort the following day but this tends to be short lived. Undertaking maximum muscle
contractions poses a low risk of muscle strain. Back injury is not expected as similar methods have
90
been used in a previous low back study with no incidence of back injury. If you have low back pain,
your condition may or may not be slightly aggravated. The skin underlying the EMG electrodes
may show some redness and minor irritation for several hours after the completion of the
experiment.
How will these discomforts and risks be alleviated?
You will be able to terminate testing at any time. You will be required to perform a 10 minute
standardised warm up which will include riding an exercise bike for five minutes followed by lifting a
box 10 times from floor to mid-thigh. To further assist warming up, you will then perform a specific
warm-up consisting of three back extensions at 50 percent of maximal effort in the different test
positions. This warm-up will help to reduce the chance of any muscle strain or soreness occurring
while making the testing procedure feels familiar.
A mandatory rest between test positions (two minutes) will also decrease the risk of strains
occurring as a result of fatigue. For additional comfort, following each maximum effort you will be
able to unclip yourself from the chain and stand upright.
I am a registered physiotherapist who is trained in cardiopulmonary resuscitation will be doing the
testing therefore I will be able to identifying potentially hazardous situations that may occur during
the testing. You will be able to communicate with me at all times during the testing procedures and
may cease participation in the study at any time. You will be given my contact details should there
be any discomfort following the testing session.
To address the potential skin irritation the skin will be cleaned before and after application of the
electrodes. Aloe vera cream will be available in the laboratory to reduce any irritation if required.
To minimise muscle soreness or strains you will be given a warm-up as outlined in ‘What will
happen in this research’. Aloe vera will also be available to sooth any minor skin irritation where the
electrodes were placed.
What are the benefits?
You will receive no direct benefit from participation in the study. The study findings will help to
determine how the different sections of the back muscles work in different ranges of lumbar flexion
and under the different conditions of pelvis fixation. Knowing what normally occurs and comparing
to case samples with people with lower back pain will provide necessary information on how back
muscle muscles are affected by injury. This in turn will help guide the rehabilitation of back injuries.
It will also benefit researchers in future investigations related to the optimal back muscle
rehabilitation methods. The researcher will benefit by gaining a Masters of Health Science.
What compensation is available for injury or negligence?
In the unlikely event of a physical injury as a result of your participation in this study, rehabilitation
and compensation for injury by accident may be available from the Accident Compensation
Corporation, providing the incident details satisfy the requirements of the law and the Corporation's
regulations.
How will my privacy be protected?
No material which could personally identify you will be used in any reports in this study. For the
analysis of the data, each participant will be given a confidential coding, so that their information
can be linked. After the analysis, the data will be kept locked in a filing cabinet in the Health and
Rehabilitation Research Institute, Auckland University of Technology. The consent forms for the
study will be stored in a similar fashion but will be kept separate from the data collected. All results
are pooled, so no names or participants or any material that could identify an individual will be
published or presented.
What are the costs of participating in this research?
There are no monetary costs involved in taking part in this study. The only cost to you is the time
in attending the testing session which will be no longer than two hours.
What opportunity do I have to consider this invitation?

91
Once you have received and read this information you will need to inform the researcher if you
intend on participating within two weeks. If you are interested in participating, you will be given a
verbal explanation of the study and any questions you have will be answered. An appointment to
attend will be made at least one week away, so that you have a further week to consider whether
you wish to take part. Please remember that your participation is voluntary and you are able to
withdraw from the study at any time by informing the researcher.
How do I agree to participate in this research?
You need to contact Laura Holder, 09 921 937 or 021 488430, laura.holder@aut.ac.nz Before
participating in this research project you will be required to sign a consent form to acknowledge that
you understand what is involved and that you are happy to participate. This will be done on the day
of testing and before testing begins.
Will I receive feedback on the results of this research?
You are given an opportunity on the Consent Form to indicate if you would like feedback on the
research project. If you answer “yes” to this, a 1-page summary of the results of the project will be
sent to the contact details that you provide on the Consent Form.
What do I do if I have concerns about this research?
Any concerns regarding the nature of this project should be notified in the first instance to the
Project Supervisor, Dr Grant Mawston, grant.mawston@aut.ac.nz tel: 09 921 9999 ext 7180
Concerns regarding the conduct of the research should be notified to the Executive Secretary,
AUTEC, Madeline Banda, madeline.banda@aut.ac.nz , 921 9999 ext 8044.
Whom do I contact for further information about this research?
Researcher Contact Details:
Laura Holder, laura.holder@aut.ac.nz tel: 09 921 9377, mob: 021 488430
Project Supervisors Contact Details:
Dr Grant Mawston, grant.mawston@aut.ac.nz tel: 09 921 9999 ext 7180
Associate Professor Mark Boocock, mark.boocock@aut.ac.nz tel: 09 921 9999 ext 7167

Approved by the Auckland University of Technology Ethics Committee on 19th April 2011, AUTEC Reference
number 11 / 15.

92
Appendix D

Consent Form

Project title: The effect of lumbar spine position on back extensor torque and trunk
muscle activation.

Project Supervisor: Dr Grant Mawston.

Researcher: Laura Holder.
 I have read and understood the information provided about this research project in the
Information Sheet dated 05.07.2011.
 I have had an opportunity to ask questions and to have them answered.
 I understand that I may withdraw myself or any information that I have provided for this
project at any time prior to completion of data collection, without being disadvantaged in any
way.
 I am not suffering from heart disease, high blood pressure, any respiratory condition (mild
asthma excluded), any infection, any illness or injury, other than back pain, that impairs my
physical performance.
 I am not pregnant.
 I agree to take part in this research.
 I wish to receive a copy of the report from the research (please tick one):
Yes No

Participant’s signature:
.....................................................…………………………………………………………
Participant’s name:
.....................................................…………………………………………………………
Participant’s Contact Details (if appropriate):
………………………………………………………………………………………..
………………………………………………………………………………………..
………………………………………………………………………………………..
………………………………………………………………………………………..
Date:
Approved by the Auckland University of Technology Ethics Committee on 19 th April 2011
reference number 11 / 15.
Note: The Participant should retain a copy of this form.

93