Stretching and Low Back Pain
Abstract
The etiology of mechanical low back pain is far from certain, and so causative links with stretching are speculative. The widely accepted argument attributing back pain to actual damage to the spinal structures such as the discs and facet joints has been strongly challenged by Waddel (1998) who argues that back pain may be largely attributed to soft tissue dysfunction strain caused through misuse and poor posture but without structural damage. This accords with recent significant research linking low back pain with spinal instability (Panjabi 1992) and this has led to the emergence of new remedial techniques aimed at promoting spinal stability (Richardson et al 1999). These developments enhance the need of a critical review of the role of stretching soft tissues in the context of exercise and therapy as flexibility remains a poor predictor of future back pain (Sullivan et al 2000).
Introduction:
The postures and exercises considered in this series involve elongation of soft tissue that induces tensile stress and, with the exception of traction, also impose compression of soft tissue. The effect of these forces on the microstructure of the soft tissues was considered in part 1 of this series. This final part considers the effect of these forces on the overall spinal structure. The optimal level of balance between strength and flexibility is a subject of conjecture between different biomechanical models of the spine. Predictions of the effects of stretching and posture work will necessarily vary in accordance with the biomechanical model of the spine on which predictions are based. Key arguments concern the imposition of pressure on the intervertebral discs that is interpreted differently according to different theoretical models exemplified respectively by Kapanji in 1974 and Aspden in 1996. Parallel to these models is conjecture of the relative contributions to back pain from structural defects and postural dysfunction that stretching exercises may help to promote or alleviate (Waddell 1998). This decrees that the cause of most back pain is due to movement disorder and muscle inefficiency but without structural damage. The corollary of this is the necessity of control and stability by the local muscles, in particular the multifidus and transverse abdominis. Excess flexibility and lack of muscle tone may compromise stability of the lumbar spine. This need not herald the demise of disciplines involving stretching of soft tissues as spinal stabilisation muscles can be trained whilst performing stretching exercises if there is simultaneous abdominal and pelvic floor muscle contraction. Exercise regimes suitable for relieving and preventing mechanical low-back pain therefore need to incorporate postural and strength training combined with flexibility training to balance lordosis and facilitate normal range of movement.
Keywords: Mechanical
low back pain, stretching, biomechanics
Methodology
Searches were conducted
for reports of studies on the effects of stretching soft tissues generally and
those of the low back in particular. Reports of random controlled trials
relating stretching to relief or inducement of lower back pain were of primary
concern. Reports on the effects of stretching on injury prevention and athletic
performance were also considered.
a. References to respective trails and
studies were derived from published meta reviews. This report reviews only
pertinent trials and studies that the meta reviews considered to be of good
quality.
b. Texts on the biomechanics of the low back
and the biomechanics of stretching soft tissues were sought from major
contemporary authors.
c. Texts were sought describing the practices
of stretching tissues of the low back used by manual therapists and fitness,
yoga and pilates instructors.
Literature
searches were conducted on the British Library catalogue and the internet using
Medline, Pubmed, Science Direct, BMJ.com and other search engines. The review
aimed at a conceptual qualitative analysis.
Literature
review
The relative flexibility and strength of soft tissues have a direct
influence on the curve of the lordosis, intervertebral disc pressure and spinal
stability. The exact influence of stretching is subject to conjecture manifest
in different biomechanical models. Kapandji (1974) offers an analogy of the
spine as a ship’s mast resting on the pelvis and extending to the head held in
place at all levels by muscles and ligaments acting as ‘stays’. When opposing
forces are in equilibrium the spine is held in correct alignment and curvature.
Any relative weakness or stiffness on one side pulls the spine out of alignment
causing stress and possible pain. Unduly stiff parts of the spine are
compensated by other parts of the spine working harder to maintain overall
spinal mobility incurring wear and tear on spinal structures leading to overuse
injuries. Muscles and ligaments determine the shape of the lordosis subject to
the constraints of the bony vertebrae.
Kapanji (1974)
explains the forces imposed on the lumber intervertebral discs in spinal
flexion as the product of force induced by the abdominal muscles acting as a
lever that initiates flexion with the fulcrum at L5-S1 and the erector spinae
serving as counter-levers to control movement. This model is generally used for
ergonomic calculations but the logic extends to forces imposed on the spine by
the weight of the head and upper body in forward flexion.
Aspden (1996)
argues that the lever model of the spine is over-simplistic, conferring little
reference to the engineering features of the spinal curvature. Aspden (1996)
argues that the forces on the spine calculated from the model almost equal the
failure strength of the intervertebral discs. The model was improved by
incorporating intra-abdominal pressure as a factor providing a counter force to
forward moment. The anterior pressure on the spine also causes it to lengthen,
thereby spreading the vertebrae and relieving pressure on the discs (Corrigan
and Maitland 1998).
However, Aspden (1996) argues that even accounting for the contribution
of intra-abdominal pressure the spine in this model would still be working to
physiological limits carrying even modest weights.
Aspden argued the
role of abdominal muscles in protecting the spine is by imposing compression
that helps to stiffen the spinal structure. This is couched in terms of his
model of the spine as an arch deriving stability from compression imposed by
muscles and ligaments intrinsic to its structure.
Longitudinal tensile forces coupled with tensile forces against the
concave surface stabilise an arched structure in the manner of a suspension
bridge. Longitudinal compression coupled with compression against the convex
surface confers stability on an arched spine in the same manner as an arched
masonry bridge. Circular ‘classical
Roman’ arches require less compression to maintain stability than oval ‘Gothic’
arches but Gothic arches have greater resistance to force. In the lumbar spine,
the more circular excessive lordosis curve is likely to be less resistant to
compression (Aspden 1996).
An arched structure
can carry a load without the need for an external balancing moment. Stability
is conferred on an arch if the imposed forces are transmitted through a
‘thrust-line’ trajectory passing longitudinally within the structure. This
trajectory is analogous to ligaments and muscles that pull vertebral segments
together. Stability to an arched structure requires compressive forces to be
developed around its curves. Forces are
transmitted through a thrust line that in a stable arch runs through the
internal length of an arch. In the
spine, muscles and ligaments can provide the necessary intrinsic support if
positioned close to the spinal axis and can maintain force throughout spinal
movements. The intradiscal pressure provided by the ligamentum flavum and the
longitudinal ligaments (Nachemson & Evans 1968; Rolander 1966 – both cited
in Nordin and Frankel 2001, p260) concur with this model. The longitudinal
ligaments attached to the vertebral bodies
maintain compressive forces throughout flexion and extension. Increased
curvature increases the stiffness of the ligaments and their thrust on the arch
enhancing its stability. The influence of the ligaments is at the outer range
of vertebral segmental movement. The paraspinal muscles provide control at the
inner range (Panjabi 1992b).
The main difference between an arch
and a lever is that a lever needs external support whilst an arch is
intrinsically stable provided there is sufficient compressive thrust. The
balance of strength and flexibility of the muscles and ligaments are therefore
of a different calculation in the arch model than appropriate to the ‘mast’ or
‘lever’ model and the role of stretching and strengthening changes
accordingly. The force muscles need to
generate in the ‘lever’ model requires them to be positioned on a longer lever
facilitated by distance from the spinal axis. In the lever model the spine will
obtain more protection if the erector spinae has been lengthened by stretching
because it will exert optimum counter moment force at longer lengths
corresponding with greater degrees of forward flexion. In the arch model
compression exerted by muscle contraction mostly at its inner range serves to
stabilize the spine. The influence of muscle adaptation and elongation on the
ability of spinal muscle to exert force at its inner range is therefore crucial
to the stability of the vertebral joints.
Muscle fibres have been shown in
vitro to adapt to chronic stretching by producing additional sarcomeres. This provides an increase in the cross
sectional area of the muscle that increases the muscle force potential (McComas
1996; Lieber 1992). Contractile force is enhanced as more actin and myosin
cross bridges can remain linked as the muscle is elongated. Muscles are thereby
able to produce a given force at a greater length. Stretching erector spinae to
the point of inducing additional sarcomeres could therefore increase their
contractile force and their ability to control spinal flexion. However, an
adaptively elongated muscle has to accommodate its additional sarcomeres within
the original shortened length if it is to retain the same contractile force in
its inner range of motion. Contractile force may be reduced if the actin and
myosin fibres become cramped and overlap. The compression required for spinal
stabilization in the arch model is derived from the inner range force of the
muscle. Stretching multifidus to the
extent of generating additional sarcomeres may therefore compromise the
compressive forces necessary to protect and stabilise the spine (Norris 2000).
A similar fate is likely if the transversus abdominis is stretched in back
extension postures making it less able to create and maintain intra abdominal
pressure and contribute to the stabilisation of vertebral motion segments.
Structural damage to the intervertebral disc and facet joints as a major
cause of back pain has long been widely accepted and reported in classic texts
including Cyriax (1993), Kapandji (1974) and Corrigan and Maitland (1998). The function of the discs and facet joints
are inter-related as is any ensuing pathology. The forces imposed on the discs
and facet joints are subject to the curvature of the lumbar lordosis that is
influenced by the relative strength and flexibility of the associated muscles
and ligaments attached to the lumbar spine and the pelvic girdle (Kapanji
1974). These forces influence pain or its relief (Adams et al 2002) and are in
turn directly influenced by stretching and strengthening exercises.
Disc degeneration may result in back pain as the discs become less
flexible and less able to store fluid and resist compression (Bogduk and Twomey
1991). This may accelerate if excessive or wear and tear accrue with long-term
misalignment of movement. Loss of disc height deranges the facet joints,
causing strain to the fibrous tissues of the joints and increasing potential
for arthrosis. The spaces between
adjoining vertebrae through which spinal nerves exit the spinal cord are
narrowed as the discs lose height. This causes surrounding structures,
including facet joints, to press against the nerves that innervate the hips and
lower limb, causing sciatica or other local or referred pain. The facet joints
link the vertebral bodies and are located posterior to the discs. The joints are
enclosed in fibrous capsules, layered with cartilage that absorbs compression.
The joints secrete synovial fluid that lubricates movement and provides a
vacuum that pulls together the articulating surfaces. These features are
compromised if the joint is damaged.
Facet joints can degenerate independently of the disc and be a source of
low back pain often precipitated by minor trauma. However, most facet joint
problems are associated with disc degeneration where loss of disc space causes
joint impingement. The facet joints normally protect the disc from rotational
stress but cannot do so when impacted with loss of disc height. This in turn
causes compression and facet joint degeneration (Bogduk and Twomey 1991).
Disc
disruption is attributed to 40% of reported cases of mechanical low back pain
and small fractures and tears of the facet joints account for 10%-15% of cases.
The sacroiliac joints are implicated in about 20% of cases of mechanical low
back pain although the pathology remains unknown. Spondylolysis is often
asymptomatic although it may cause back pain particularly in athletes. Data is
lacking on the diagnosis and prevalence of dural pain and there are
difficulties in diagnosing ligament and muscle sprains (Adams et al 2002). Waddell (1998) challenges the correlation between structural damage and
low back pain by reference to X-rays and CAT scans showing impaired spinal
structure of persons claiming to be free of pain. He argues that disc prolapse
is rare and discs degenerate with age without necessarily inducing back pain
and notes that 60% of adults have back pain each year whilst only three to five per
cent have ever had a disc prolapse in their lives. Furthermore, MRI scans
provide consistent evidence of worn discs and faulty spinal structure but
without accompanying back pain. Conversely, scans
reveal normal musculoskeletal structure of persons claiming to suffer from
chronic back pain where there is no evidence of pathology. Waddell also doubts
the links between facet joint injury and back pain. He cites Jackson et al (1988) who
demonstrated facet joint pain occurring in back extension was not relieved by
anesthetic injection; Lilius (1989) in controlled trials found injections into
and around the facet joints no better than a placebo; van Tulder et al (1997)
found no link between X-ray depictions of degenerative changes in the facet
joints and back pain.
Waddell (1998) argues that dysfunction rather than structural damage is the cause of most back
pain. He defines dysfunction as a self-perpetuating syndrome of movement
disorder and muscle inefficiency. This may be associated with physical stress,
strain, increased or unaccustomed use, fatigue, poor fitness and poor posture.
Dysfunction resulting from poor posture accords with evidence, cited by
Corrigan and Maitland (1998), that chronic postural stress asymmetrically
imposed on soft tissues causes an increase in fibroblast activity and
subsequent collagen production. The additional fibres encroach on space in the
connective tissue normally occupied by nerves and vessels. This can cause a
loss in muscle elasticity and may result in pain on muscle activity. Postural
muscles become stronger but tighter whilst the ‘phasic’ muscles that control
more rapid movement become weaker and longer. Postural muscles therefore need
to be stretched and phasic muscles strengthened in the context of postural
training in the process of treating back pain associated with ‘dysfunction’.
Panjabi (1992a) focused on spinal instability as an important cause of
low back pain whereby inappropriately large intervertebral movements cause
compression or stretching that abnormally deform soft tissues. Panjabi (1992b)
argues that spinal instability occurs at the ‘neutral zone’ of intervertebral motion
where little resistance is offered by the ligaments. The neutral zone becomes
larger with injury or muscle weakness and smaller with increased muscle force.
Control and stability at the neutral zone are by the local muscles of the
lumbo-pelvic area with multifidus having a particularly important role. These
deep local muscles have their insertions or origins attached to the lumbar
vertebrae and control inter-segmental motion and local stabilization. Larger
muscles situated more laterally provide ‘global’ stability in the back. The global muscles handle the external loads
placed on the trunk, so minimizing the variations in load imposed on the local
muscles to manageable levels The ligaments and other passive elements limit
motion only at the outer ranges of joint movement but their role as
proprioceptors influence local muscle movements (Richardson et al 1999).
Arguing from the basis of
the Panjabi (1992a; 1992b) studies, Richardson et al (1999) agree with Waddell
(1998) that dysfuntion is a significant cause of mechanical low back pain. In
so doing, they highlight the role of the local muscle system, particularly
transversus abdominis and multifidus, Multifidus is an important spinal
stabilizer that functions in close coordination with transversus abdominis, the
pelvic floor muscles and the respiratory diaphragm (Richardson et al 1999).
Multifidus may be stimulated directly but its activity is coordinated with the
stimulation and contractions of transversus abdominis. Multifidus has
considerable influence on lumbar segmental stability linking vertebrae to
vertebrae within the lumbar and between lumbar and sacral vertebrae. It finely
adjusts spinal movement and enhances stability by stiffening motion segments.
Facet joints are covered by multifidus on all sides except where the joints are
in direct contact with ligamentus flavum so the multifidus muscles keep the
joint capsules taut and prevent impingement of articular cartilage. The
functional activities of multifidus is linked to transversus abdominis with
both acting as synergists as integral components of a corset that imposes
intra-abdominal pressure (Richardson et al 1999). Transversus abdominus in turn
has considerable influence on the tension of the thoracolumbar fascia
facilitated by its extensive attachments to it. The thoracolumbar fascia is
thought to constrain radial expansion of the multifidus and the lumbar
longissimus and iliocostalis muscles (Aspden 1996). This provides a brace around these muscles that enhances their
strength contributing to spinal stiffness and lumbar stabilization (Gracovetsky
et al 1977 cited in Richardson et al 1999). Excessive stretching of the fascia
may therefore result in loss of muscle strength and spinal stability.
Evidence of dysfunctional
multifidus in low back pain patients is demonstrated in four areas. Firstly,
there is evidence of atrophy with muscle fibres containing a larger percentage
of fat (Cooper et al 1992, Mooney et al 1997 both cited in Waddell 1998).
Secondly, there is evidence of decreased electro myograph activity at the
unstable level during concentric back activity (Sihvonen et al 1991 cited in
Richardson et al 1999, p69). Thirdly, the muscle has been found to fatigue more
easily (Biedermann et al 1991; Roy et al 1989 both cited in Richardson et al
1999, p69). Finally, abnormal degenerative changes in internal muscle fibre
structure have been found in low back pain patients (Matilla et al 1986 cited
in Richardson et al 1999 p 70). Subjects with low back pain have been found to
have significant delay in the onset of contraction of transversus abdominis in
preparation for movement of the limbs compared to control subjects. This is not
necessary associated with weakness of the muscle but suggests inadequacy in
motor control (Richardson et al 1999).
Richardson et al (1999)
argue that the link between muscle function and spinal stiffness provide a
basis for therapeutic exercise to be used to manage spinal instability,
although they concede more detail is needed about the muscle and related neural
control system. Nevertheless, these authors offer a specific exercise strategy
based on co-contraction of the multifidus, transversus abdominis, the pelvic
floor muscles and respiratory diaphragm aimed at promoting segmental
stabilization to treat mechanical low back pain. The co-contraction techniques
described by Richardson et al (1999) are conducted initially in isolation but
are progressively incorporated into other postural activity. Interestingly, the
abdominal and pelvic floor muscle contraction techniques are a distinct
practice in yoga (Coultner 2001) and are incorporated in the practice of
Astanga yoga postures (Bender Birch 1995) similar to those illustrated in part
3 of this series.
Conclusion:
The overriding consensus from the literature and those interviewed for
this series was that stretching should be felt but free from pain. Essentially,
those stretching need to respect the pain response evolved to prevent tissue
damage whilst creating enough force to produce the desired result. This is not infallible
as ‘stretch tolerance’ (Shrier 1999) can develop as those stretching regularly
become less sensitive to neural responses. Maitland et al (2001) offers a model
for manual therapy stipulating the need to marry knowledge of structure,
function and pathology with empirical observation of the patient. Treatment is
based on pain and sensory response requiring good communication between
instructor and client. This approach could be applied to stretching, whereby
the trainer guides the client in recognising and interpreting sensory responses
and modifies the stretch accordingly. Biomechanical knowledge of the
capabilities and physical limitations of soft tissues may help to guide
interpretations of the sensory response.
In the longer term, controlled trials isolating the effects of specific
stretching exercises would contribute to clinical decisions on their
employment.