Clinical Biomechanics 1987; 2: 78-83 Printed in Great Britain

The biomechanics of the thoracolumbar

fascia
Janet E Macintosh BSc*
Nikolai Bogduk MB BS*, Serge Gracovetsky Pm*

*Department of Anatomy, University of Queensland, St Lucia 4067, Queensland, Australia

*Department of Electrical Engineering, Concordia University, 1455 de Maisonneuve Blvd. West, Montreal H3G 1 M8,
Quebec, Canada

Summary
The back muscles alone are unable to provide the extensor moment required to lift large
weights, and must be aided by another source of anti-flexion moments. It has been
postulated that contraction of the abdominal muscles can provide an extension moment by
developing tension in the thoracolumbar fascia (TLF). Anatomical studies and a biomecha-
nical analysis, however, reveal that the anti-flexion moment generated in this way is only
very small. Too little of the abdominal musculature attaches to the TLF to generate a
significant tension in it. Previous calculations of the forces in the TLF have overestimated
the tension developed in it because of erroneous assumptions and interpretations of the
relevant anatomy. Whatever the role played by the TLF in lifting it must be essentially
independent of abdominal mechanisms.

Relevance
This study illustrates the importance of consulting or determining the anatomy integral to a
biomechanical theory before undertaking calculations and ascribing the functional signi-
ficance of postulated mechanisms. Controversies concerning tne possible mechanisms that
assist the back muscles during the act of lifting are addressed.

Key Words: Lumbosacral region, Thoracolumbar fascia, Anatomy, Biomechanics, Lifting, Abdominal
muscles, Intra-abdominal pressure

Introduction lift ii weight. the back musclc~ must bc assist4 by an

additional mechanism that ;IISO contributes iI backward
In biomechanical terms. the cardinal process involved
bending moment.
in lifting is the balancing of moments. Gravity acting on
The citrlicst theory that dcscribcs such iI mechanism
the trunk and any weight to be lifted exerts a forward
is the iihdominiIl hitllo(Jn theory Al’ Bartclinki. The
bending moment on the trunk. To raise the trunk. and
essence of this theory is that contraction of the ahdo-
thereby the weight. this moment must be b:ll;knccd and
minal muscles produces it rise in intra-abdominal press-
overcome by a backward bending moment. and such a
ure which effectively creates ;I ‘bitllo()n in front of the
moment can be provided by the back muscles acting
vertebral column that resists compression and. there-
posterior to the axes of sagittal rotation of the lumbar
fore, forward bending. Forward hcnding moments are
vertebrae.
thereby halanccd by iI combination of back muscle
In the absence of any external load. the back muscles
activity irnd raised intra-nbdominal pressure.
are sufficiently strong to raise the trunk from a forward
Although this theory hiis been popularized and SUP-
bent position. However, various calculations have
ported by circumstantial evidence. recent studies have
shown that the back muscles alone are insufficient to
yielded incompatible or pitrild(>xiciil data. Abdominal
raise the trunk when substantial external loads are to be
muscle activity does incrcasc during lifting in propor-
lifted’-‘. Their contraction is unable to overcome the
tion to the weight lifted or the Hcxion moment gener-
combined forward bending moment generated by the
ated ‘V’3. However. while the intru-abdominal pressure
weight of the trunk and the large weight to be lifted. To
correlates well with the load on the spine when it is
statically loaded, it correlates poorly with the dynamic
loading and is influenced by postural asymctry”-I”.
Moreover, the large initial increase in intra-abdominal
pressure that occurs during ;Imaximum lift is only brief.
 Macintosh et al.: Biomechanics of the thoracolumbar fascia 79

and lower pressures are present throughout most of the an anti-flexion moment. This tension in the thoraco-
lift’.17 lumbar fascia could be generated by contraction of the
Exercises designed to increase the capacity of the abdominal muscles that attach to the thoracolumbar
abdominal muscles to raise intra-abdominal pressure fascia and by the expansion of the back muscles that
succeed in strengthening the muscles. but do not result underlie the fascia. The latter process was described as
in higher intra-abdominal pressures during liftinglx.‘“. the ‘hydraulic amplifier mechanism”x.
Although patients with back pain have reduced abdo- These proposals, however, were made in the absence
minal muscle strength. their intra-abdominal pressure of any appropriately detailed anatomical data on the
during lifting does not differ from that in controls”‘, thoracolumbar fascia. Having recently completed a
and when their abdominal muscles are retrained and study of the thoracolumbar fascia”’ we are now able to
strengthened. intra-abdominal pressure does not offer both a qualitative and quantitative assessment of
?I the posterior layer of thoracolumbar fascia and its
improve.
On theoretical grounds. it has been calculated” that putative rote in lifting.
to execute a lift of 100 kg. an intra-abdominal pressure
of 2SOmmHg (32 kPa) is required to balance the bend-
Anatomy
ing moment. Were this level of pressure to be sustained
for longer than a brief moment the abdominal aorta The thoracolumbar fascia consists of three layers: an
would be occluded and blood flow to the viscera and anterior layer which arises from the anterior surface of
lower extremities would be compromised (which in- the lumbar transverse processes and covers quadratus
deed. is a reservation raised by Bartelink himself’). lumborum: a middle layer. from the tips of the lumbar
Furthermore. the known capacity of the abdominal transverse processes; and a posterior layer which arises
muscles is 60--5Opsi (0.4-14 MPa)“.‘” which is insuf- from the midline to cover the back muscles’“.
ticient to realize the level of hoop tension required to The posterior layer of thoracolumbar fascia covers
generate such pressure. the back muscles from the lumbosacral region through
Not withstanding these quantitative limitations. in- to the thoracic region as far rostrally as the splenius
tra-abdominal pressure is raised by contraction of the muscle. In the lumbar region it is attached to the
external and internal abdominal obliques and to a midline. and lateral to the erector spinae. between the
lesser extent rectus abdominis”. but it has been noted 12th rib and the iliac crest. it unites with the middle
that these muscles also exert a flexion moment on the layer of thoracolumbar fascia forming a raphe. referred
trunk, which tends to negate any antitlexion effect to as the ‘lateral raphe”“. At sacral levels. the posterior
exerted by the raised intra-abdominal pressure”. la!cr cstcnds from the midline to the posterior superior
These several limitations and inconsistencies ha\,c iliac hpine and the posterior segment of the iliac crest.
brought the intra-abdominal balloon thcor!, into ques- At lumbosacral levels. the posterior layer consists of
tion and have led to the proposal of another mechanism tl\\o Iaminae: a supcrticial lamina with hbres orientated
capable of producing an extensor force. as an altcrna- ciltIdomcdi;lll~ and ;I deep lamina with tibres oriented
tive or in addition to the abdominal balloon mcchan- caudolatcral~y (Figure 1). The superficial lamina is
ism. Gracovetksy. Farfan and Lar-ny”’ proposed that t’orrncd b!, the aponeurosis of latissimus dorsi. but the
the posterior ligamentous system is capable of storing disposition and attachments of its constituent fibres
sufficient tension to enable the vertebral column to vilr!‘. The most lateral fibres from latissimus dorsi
overcome any flexion force that exceeds the capacit\ of atti\ch \,ia short tendons directly to the iliac crest and do
the back muscles. The posterior ligamentous system not contribute to the posterior layer of the thoracolum-
consists of the capsule of the zygapophysial joints and bar i’ascia. Marc medial fibrcs become aponeurotic just
the midline ligaments. and the posterior layer of the lateral to the lateral raphc. As they pass through the
thoracolumbar fascia. raphc the tendons are deflected medially so that they
The tensile properties of the zygapophysial capsules eventually gain attachment to the lower lumbar and
and the supraspinous. interspinous and Haval ligaments sacral spinous processes. forming the sacral portion of
have been studied previouslyZ7. and collectivcl! the! the supcrticial lamina of the posterior layer. The most
can balance between 24% and 55% of applied Hcxion medial tibres pass through the lateral raphe and gain
moments. This action. howcvcr. depends 011 the attachment to L3, L4 and LS spinous processes. and
amount of lumbar tlexion. for it is necessar!’ for the form the lumbar portion of the superficial Iamina of the
spine to elongate posteriorly. and remain csscntiallq posterior layer.
fully flexed. for tension to be stored in the ligaments. The deep lamina consists of bands of collagen fibres
Gracovetsky et al.‘s and later Gracovetsky ct al.’ emanating from the midline, principally from the lum-
proposed that the thoracolumbar fascia could provide bar spinous processes (Figure I). The bands from the
an additional anti-flexion moment that is not dependent L3. L.‘, and Sl spinous processes pass caudolaterally to
on the spine being fully flexed. It was suggested that. the posterior superior iliac spine. Those from the L3
because the fibres of the posterior layer of thoracolum- spinous process and L3-4 interspinous ligament wrap
bar fascia attached obliquely to the lumbar spinous around the lateral margin of the erector spinae to fuse
processes, tension in the fascia could act to prevent with the middle layer of thoracolumbar fascia in the
separation of the spinous processes. thereby providing lateral raphe. Above L3 the deep lamina progressively
 80 C/in. Biomech. 1987; 2: No 2

Figure 1. A posterior view of the posterior layer of

Figure2. From any point in the lateral raphe (Ir) two
thoracolumbar fascia. On the right, the deep lamina
collagen fibres diverge towards the midline through the
alone is shown, with bundles of collagen fibres passing
posterior layer of thoracolumbar fascia. One passes
from the L4, L5 and Sl spinous processes to the posterior
upward in the deep lamina, the other downwards in the
superior iliac spine (PSIS), and from L2 and L3 to the
superficial lamina. The divergence between the two
lateral raphe (Ir) where they fuse with the aponeurosis
fibres spans two vertebral levels.
(apon) of transversus abdominis (TA) which forms the
middle layer of thoracolumbar fascia. On the left, the
deep lamina is covered by the superficial lamina which is
formed by the aponeurosis of latissimus dorsi (LD) and of the triangle (Figure 3). Because of the obliquity of
which fuses with the deep lamina and the middle layer of the borders, the force exerted at each basal angle would
thoracolumbar fascia in the lateral raphe. The two consist of two components: a horizontal and a vertical
laminae endow the intact posterior layer of vector (Figure 3). When lateral tension is applied to the
thoracolumbar fascia with a cross-hatched appearance.
thoracolumbar fascia bilaterally the sum of the horizon-
tal vector forces at the midline will be zero. but the
remaining vertical vectors in any triangle will be ex-
becomes thinner, consisting of sparse bands of collagen erted in mutually opposite directions. This arrange-
that dissipate laterally over the erector spinae. A deep ment allows lateral tension in the posterior layer to
lamina is not formed at thoracic levels. exert a force that tends to approximate the spinous
Collectively, the superficial and deep laminae of the processes, and is equivalent to an extending. or anti-
posterior layer of thoracolumbar fascia form a retinacu- flexion force.
lum over the back muscles. Attached to the midline The middle fibres of the transversus abdominis attach
medially and the posterior superior iliac spine and to the lateral raphe of thoracolumbar fascia”‘. and
lateral raphe laterally, the fascia covers or ensheaths therefore their contraction is capable of generating
the back muscles, preventing their displacement dorsal- lateral tension in the posterior layer. Apart from a few
ly. However, recognition of the disposition of its com- fibres of internal oblique that may attach to the
ponent fibres allows the mechanical properties of the raphe”‘, transversus abdominis is the only abdominal
posterior layer to be discerned.
The obliquely crossing fibres of the two laminae of
the posterior layer are arranged such that at any given
point on the lateral raphe a fibre from the superficial
lamina extends rostromedially. The posterior layer
therefore consists of a series of overlapping triangles of
collagen fibres whose apices lies in the lateral raphe and
whose bases lie in the midline (Figure 2). The upper
border of each triangle is formed by a fibre in the deep
lamina, and the lower border is formed by a fibre in the
superficial lamina. The divergence of the two borders is
such that each triangle subtends two vertebral levels.
When the lumbar spine is in the neutral position, the
fibres of the posterior layer are orientated at 30 degrees
to the horizontal, but in the flexed position this angle
increases to about 40 degrees.
Because of the triangular arrangement of its fibres, Figure 3. Lateral tension applied to the posterior layer of
lateral tension forces applied to the posterior layer will thoracolumbar fascia will be distributed in a triangular
fashion towards the midline. The divergence of the fibres
be transmitted in a triangular fashion. Lateral tension of the posterior layer permits this lateral tension to exert
applied to the apex of any triangle in the lateral raphe an extending force on the lumbar spinous processes (see
would be transmitted to the midline along the borders Figure 4).
 Macintosh et al.: Biomechanics of the thoracolumbar fascia 81

muscle that attaches to the posterior layer, and what-

ever influence the abdominal muscles may have on the
thoracolumbar fascia will be expressed only by the
transversus abdominis. Knowledge of the detailed ana-
tomy of the thoracolumbar fascia makes it possible to
calculate the magnitude of the tension generated by the
transversus abdominis and the size of any anti-flexion
moment generated by it.

Calculations

Contraction of transversus abdominis would apply

lateral tension (TL) to the lateral raphe, which would
be transmitted to the midline through fibres of the
posterior layer of thoracolumbar fascia. If the fibres of Figure4. The geometry of tension forces (TJ exerted on
the posterior layer are orientated at 0 degrees to the the midline by the lateral tension (TJ applied to the
horizontal, the tension that develops in each fibre (Tr) fibres of the posterior layer of thoracolumbar fascia
which are orientated at 0 degrees to the horizontal. r,
can be resolved into a horizontal tension (T,,) and a
and TV represent the horizontal and vertical vectors of Tf.
vertical tension (7’“) (Figure 4):

T,,=T, cos 8 (1) area of that part of the transversus that acts on the
posterior layer is 14mm”.
Tv=Tf sin 9 (2) Assuming maximum contraction and a generous
coefficient of contraction of 10 kg/cm’ (1 N/mm’)3”, the
TL must equal the sums of the horizontal tensions in tension (TL) exerted by the two transversus muscles
the triangles ABC and ADC. i.e. T,_ = 2Th. would be:
From (1,:
Tl_ = 1.4 x 1Okg
T,_=‘T,.cos H = 13kg
From equation (5), the maximum extension force
TL (EF) exerted by the transversus abdominis would be:
Tf=------ (3)
3 cos H
EF = 14 tan 8 [kg]
From (2) and (3):
Since this force acts on the spinous processes whose
Tl tips lie about 5 cm behind the axes of sagittal rotation.
T,= -.sin ti
i3 cos H the extension moment (EM) exerted by the transversus
would be:
T,=$.tan 8 (J)
EM = EF x 0.05 [ kgmj
The extension force (EF) approximating B and D, = 11 tanti x 0.05 [ kgm]
exerted by the bilateral application of tension CT,_) to = 0.7 tantl [kg m]
the thoracolumbar facia will be:
During erect standing the orientation of fibres in the
EF=2T, posterior layer is 30”, with respect to the horizontal.
Therefore:
From (4):
EM = 0.7 tan 30 [kg m]
T,..tan 8 = 0.7 x 0.577 [kg m]
EF=2.
2 = 0.4 [kg m]
= 3eYNm
EF= TI_.tan Cl (5)
When the vertebral column is fully flexed. c) is 40”.
The magnitude of T,_ can be estimated from the Therefore:
cross-sectional area of transversus abdominis. To deter-
mine this the muscle was examined in ten embalmed EM = O-7 tan 40 [kg m]
human adult cadavers. = O-7 x 0+3Y [kg m]
The transversus abdominis had a mean thickness of
2 mm over the 7 cm length along which it attached to = 0.6 ]kg m]
the lateral raphe. Therefore the mean cross-sectional = 5.YNm
82 Clin. Biomech. 1987; 2: No 2

Discussion In this regard. another mechanism described to date

is the hydraulic amplifier mechanism of Farfan. Gra-
The anatomy of the posterior layer of the thoracolum- covetsky and Lamy”‘. It is conceivable that the expan-
bar fascia is such that it is capable’of transforming a sion of the back muscles as they contract could stretch
laterally applied tension into an extension moment on the posterior layer from within, raising the tension
the lumbar spine, and because the transversus abdomi- within its fibres. Using the same triangular mechanism
nis attaches to the lateral raphe of the thoracolumbar as described above, this tension could brace the lower
fascia, this muscle is capable of exerting such an exten- lumbar spinous processes and stabilize them in flexion.
sion moment. In a qualitative sense, this arrangement On the basis of the anatomy. this mechanism seems
suggests that the transversus abdominis and the thor- qualitatively feasible. However, a quantitative analysis
acolumbar fascia can act as an internal bracing mechan- still remains to be explored.
ism providing a definite extensor (anti-flexion) moment
on the lumbar spine.
Our calculations, however. reveal that the magnitude
of the force that can be exerted in this way on the erect.
References
or even fully flexed, lumbar spine, is relatively small.
Given that the maximum extensor moment generated
I Morris JM. Lucas DE. Bresler B. The role of the trunk
by the back muscles has been determined to be 2% muscles in stability of the spine. J Bone Joint Surg (Am)
28ONm”.“‘, the moment that can be generated by the 1961; 43A: 327-51
thoracolumbar fascia (3-9-5.9 Nm) represents only a 2 Farfan HF. Muscular mechanism of the lumbar spine and
small additional moment. Thus. while anatomically the position of power and efficiency. Orthop Clin North
Am lY75: 6: 135-4-l
feasible and qualitatively attractive, the abdominal
3 Farfan HF. The biomechanical advantage of lordosis and
mechanism previously described by Gracovetsky et al.’ hip extension for upright activity. Man ah compared with
provides a virtually negligible anti-flexion moment. other anthropoids. Spine 197X: 3: 3.3642
The reasons for this discrepancy are several. First. in 3 Gracovetsky S, Farfan HF. Helleur C. The abdominal
the original calculations the orientation of the fibres in mechanism. Spine 10X5; IO: 317-3-l
5 Bartelink DL. The role of abdominal pressure in
the thoracolumbar fascia was taken to be H=67” which
relieving the pressure on the lumbar intcrvcrtrbral dihca.
is greater than the observed orientation of 8=3(L40”. J Bone Joint Surg (Br) lYS7: 3YB: 71X-25
Secondly. the internal oblique was believed to contri- 6 Davis PR. Variations of the intr;i-abdominal pressure
bute to the horizontal force on the thoracolumbar during weight lifting in various postures. J Anat IYSh; YO:
fascia but anatomical dissections have now revealed 601
7 Troup JDG. Relation of lumbar spine disorder5 to he;lv!
that only transversus abdominis is effectively involved.
manual work and lifting. Lancet 1965: I : SS7-61
Moreover, only a limited portion of the transversus 8 Andersson GBT. Loads on the lumbar spine: in viva
abdominis acts on the thoracolumhar fascia. Thus the measurements and hiomech;mical analy\ez. In: Winter
original calculations over-estimated the numhcr and DA. Norman RW. Wells RP. Flaveh KC. Patla AE. 4s.
cross-sectional area of the muscles rhat could cscrt Biomechanics IX-B. Internation;; Serlez on
Biomechanics. Human Kinetic\. Champaign lYX3: 23-37
lateral tension on the thoracolumbar fascia. Com-
Y Andersson GBJ. Ortengren R. Nachmson A.
pounded, these various anatomical crror5 introduced Intradiskal pressure. intr~t-;lbtfominal pressure and
excessive multiplication factors into the calculation of myoelectric back muscle activity rclatcd to posture and
the forces. loading. Clin Orthop lY77: 13: IS&h-l
Having reappraised the biomechanics of the post- 10 Davis PR. Posture of the trunk during the lifting of
weights. Br Med J IYSY: 1: X7-Y
erior layer of thoracolumbar fascia in rhe light of new
II Davis PR. Troup JDG. Pressures in the trunk cavities
anatomical data, we are obliged to conclude that de- when pulling. pushing and lifting. Ergonomics 1964; 7:
spite that suitable arrangement of the fibrcs of the 465-74
posterior layer, too few fibres of the transversus abdo- 17 Troup JDG. Biomechanics of the vertebral column.
minis act on the thoracolumbar fascia for it to exert a Physiotherapy lY7Y: 65: 238-4-l
13 Grew ND. Intro-abdominal pressure response to loads
significant anti-flexion the role of
moment. Whatever
applied to the torso in normal subjects. Spine 19X0: 5:
the thoracolumbar fascia in lifting. it cannot bc due Id’)-54
substantially to the contraction of transversus abdo- I4 Troup JDG. Leskinen TPJ. Stalhammar HR. Kuorinka
minis. IAA. Comparison of intra-abdominal pressure-
While detiying a significant role for the transversus increases. hip torque and lumbar vertehrul compression
in different lifting techniques. tlum Factors lYX3: 25:
abdominis in the control of flexion. the present results
5 17-2s
do not exclude other mechanisms whereby the thoraco- 1s Leskinen TPJ, Stalhammar I IR. Kuorinka IAA. Troup
lumbar fascia could exert an antiflexion moment. Our JDG. Hip torque. lumbosncral compression. and intra-
studies show only that transversus abdominis is too abdominal pressure in lifting and lowering tasks. In:
small to generate sufficient tension in the posterior Winter DA, Norman RW. Wells RP. Hayes KC, Patla
AE. eds. Biomechanics IX-B. International Series on
layer. The arrangement of the fibres in the posterior Biomechanics. Human Kinetics. Champaign IYX3; 55-Y
layer nevertheless would still allow it to exert an anti- 16 Schultz AB. Andersson GBT. Ortengrcn R. Hadespeck
flexion moment, provided sufficiently high tensions K, Nachemson A. 19x3. Loads on the lumbar spine.
were generated in it. J. Bone Joint Surg. 64A: 713-10
 Macintosh et al.: Biomechanics of the thoracolumbar fascia 83

17 Davis PK. The use of intrn-abdominal pressure in
c\aluating stresses on the lumbar spine. Spine IYXI: 6:
Yo-2
IX Hemborg B. Moritz U. Hamberg J. Lowing H. Akesson
1. lntra-abdominal pressure and trunk muscle activity
during lifting-effect of abdominal muscle training in
healthy subjects. Sand J Rehabil Med lY83: IS: 18.3-06
IY Lepg SJ. The effects of abdominal muscle fatique and
training on the intra-abdominal pressure developed
during lifting. Ergonomics 1981: 24: IY I-S
30 t lemborg B. Moritz U. Intro-abdominal pressure and
trunk muscle activity during lifting. II Chronic low-back
patients. Sand J Rehabil Med IYXS: 17: S-13
21 llemborg B. Moritz U. Hamberg J. Holmstrom E.
Lowing H. Akesson I. Intra-abdominal pressure and
trunk muscle activity during lifting. 111 Effects of
;lbdominal muscle training in chronic low-back patients.
Sand J Rehabil Med IYt(S; 17: IS-24
72 Farfan HF. Gracovetsky S. The abdominal mechanism.
Paper presented at the International Society for the
Study of the Lumbar Spine Meeting. Paris. IYtil
‘3 Farfan HF. Gracovetsky S. Hellcur C. The role of
mathematical models in the assessment of tasks in the
workplace. In: Winter DA. Norman RW. Wells RP.
tlayes KC, Patla AE. eds. Biomrchanics IX-B.
International Series on Biomrchanics. fIuman Kinetics.
Champaign 1983: 3X-43
3-l Floyd WF. Silver PHS. Electromyographic study of
pattern4 of activity of the anterior abdominal wall
~nus~lesin man. J Anat lY50: X1: 132-15
2s Bean JG. The significance of the activity of the
abdominal muscles in weight lifting. Acta Anat IYhl: 4.5:
X3-Y
26 Gracovetsky S. Farfan HF. Lamy C. The mechanism of
the lumbar spine. Spine IYXI: 6: 249-Q
27 Adams MA. Hutton WC, Stott JRR. The resistance to
flexion of the lumbar intervertebral joint. Spine IYXO: 5:
245-53
2x Gracovetsky S. Farfan HF. Lamy C. A mathematical
model of the lumbar spine using an optimized system to
control muscles and ligaments. Orthop Clin North Am
lY77: 8: 135-53
2’) Bogduk N. Macintosh JE. The applied anatomy of the
thoracolumbar iascia. Spine 1984: Y: 16-l-70
30 Fick R. Anatomie und mechanik der pelenke unter
berucksichtigung der bewegenden muskeln. In: K \
Bardeleben. Handbuch der Anatomie des Menschen.
Jena: Gustav Fischer. IY 1 I
31 Gracovetsky S. Function of the spine. J Biomed Eng
1086: 8: 217-23
32 McNeil1 T. Addison R. Andersson G. Schultz AB.
Trunk strengths in attempted ilexmn. extension and
lateral bending in healthy subjects and low back patients.
Presented at the Gotesberg Meeting of the International
Society for the Study of the Lumbar Spine. lY7Y