C

TBMA

LRiTdplisad

I

Lttvsma

Lrcm

isrc[ndni

M

Iata

1

linical Review: Focused

he Role of the Lumbar Multifidus in Chronic Lowack Pain: A Reviewichael D. Freeman, PhD, MPH, DC, Mark A. Woodham, DC,

ndrew W. Woodham, BA

MvUAdgD

MD

AaCD

DC

ow back pain (LBP), a highly prevalent problem in society, is often a recurrent condition.ecent advances in the understanding of the biomechanics of LBP have highlighted the

mportance of muscular stabilization of the “neutral zone” range of motion in the low back.he lumbar multifidus muscles (LMM) are important stabilizers of this neutral zone, andysfunction in these muscles is strongly associated with LBP. The dysfunction is a result ofain inhibition from the spine, and it tends to continue even after the pain has resolved,

ikely contributing to the high recurrence rate of LBP. Persisting LMM dysfunction isdentified by atrophic replacement of multifidus muscle with fat, a condition that is besteen on magnetic resonance imaging. Muscle training directed at teaching patients toctivate their LMM is an important feature of any clinical approach to the LBP patient withemonstrated LMM dysfunction or atrophy.

PM R 2010;2:142-146

NTRODUCTION

ow back pain (LBP) is a highly prevalent problem in society; approximately 60% to 80% ofhe population will experience an episode of LBP during their lifetime, and 60% to 86% ofhese people will have more than one episode of LBP [1,2]. A relatively recent shift in theiew of LBP has gone from classifying it as a self-limited acute condition to a recurrentyndrome [3]. The purpose of this review is to discuss the role of the lumbar multifidususcles (LMM) in recurrent LBP, as well as to discuss literature-based clinical approaches to

ssessment and treatment of multifidus dysfunction.Spinal stabilization therapy has been observed to be more effective over time in treating

BP than minimal intervention and exercise therapy alone, and has also been observed toeduce pain, disability, and medication intake, as well as recurrence rates [4,5]. As aonsequence, treatment focus has shifted to reactivation and strengthening of the smalleruscles of the spine to improve long-term stabilization of the vertebral column.Biomechanical research has increased the understanding of mechanisms of low back

njury and pain, specifically regarding the ability to stabilize the “neutral zone” of the lumbarpine with tonic muscle control. Panjabi [6] has described the neutral zone as the part of theange of intervertebral motion, measured from the neutral position, in which spinal motionan occur with minimal nonmuscular passive resistance from the spine. Suni and colleagues2] described a randomized controlled trial of a neuromuscular training program directed ateutral zone stabilization on a population of patients with recent back injury. These authorsemonstrated a significant decrease in the intensity of LBP in the treatment group that wasot seen in the control group, and concluded that control of the lumbar neutral zone was an

mportant component of LBP and disability prevention.

ULTIFIDUS FUNCTION

t is well established that the LMM are important stabilizers of the lumbar neutral zone; Wilke etl [7] found that the actions of the multifidi account for more than two thirds of the stiffness ofhe spine when in the neutral zone. In comparison with all lumbar muscles, the LMM are short

nd stout, with a high cross-sectional area (CSA) and short muscle fibers. Furthermore, these

Sa

PM&R © 2010 by the American Academy of P1934-1482/10/$36.00

Printed in U.S.A.42

.D.F. Department of Public Health and Pre-entive Medicine, Oregon Health and Scienceniversity School of Medicine 1234 SW 18th

ve, Portland, OR 97205. Address correspon-ence to: M.F.; e-mail: forensictrauma@mail.comisclosure: nothing to disclose

.A.W. Private practice, Tacoma, WAisclosure: nothing to disclose

.W.W. Department of Genetic, Molecular,nd Cellular Biology, University of Southernalifornia, Los Angeles, CAisclosure: nothing to disclose

isclosure Key can be found on the Table ofontents and at www.pmrjournal.org

ubmitted for publication May 15, 2009;ccepted November 19.

hysical Medicine and RehabilitationVol. 2, 142-146, February 2010

DOI: 10.1016/j.pmrj.2009.11.006

amfiLamsmnaa

MP

AalcicraTai

paLpslcaatat

eaiitl

I

Tlr

w

fisst5do

agamnm

taLdpdd

aDposLltsa

hHasbto

RM

Mbcvtimp

143PM&R Vol. 2, Iss. 2, 2010

uthors found that at the foundation of this high CSA was a highass, which allows for packing of a large number of musclebers into relatively small space. This morphology allows theMM to produce very large forces over a small operating range,nd makes the LMM ideally suited for stability as opposed tootion [8,9]. Functionally, the LMM are divided into deep and

uperficial fibers, with deep fibers spanning 2 vertebral seg-ents and functioning tonically, and the superficial fibers span-ing 3 to 5 levels and functioning phasically [10,11]. Thisrrangement makes the deep fibers of the LMM anatomicallynd biomechanically well suited for stabilization.

ULTIFIDUS ATROPHY AND LOW BACKAIN

number of prior investigations have described the LMMtrophy and replacement by fat after low back injury, a patho-ogic process that is closely correlated with LBP. Kjaer andolleagues [12] evaluated the lumbar magnetic resonance imag-ng (MRI) results for 412 adult and 442 adolescent subjects in aross-sectional study of LMM atrophy. These authors catego-ized the degree of observed atrophy in the LMM as none, slight,nd severe, and correlated the findings with complaints of LBP.hey found that fat infiltrations of the LMM were stronglyssociated with LBP in adults, and that the association wasndependent of body mass index (BMI).

Kader et al [13] performed a retrospective study of 78atients with LBP and either with or without leg pain. Theuthors assessed the correlation between MRI changes in theMM and leg pain. It was reported that LMM atrophy wasresent in 80% of the patients with LBP and that there was aignificant correlation between LMM atrophy and referredeg pain. The authors theorized that LMM atrophy may beaused by dorsal ramus syndrome, which has been describeds LBP with referred leg pain, produced by irritation ofnatomic structures supplied by the dorsal ramus nerve, ie,he facet joints and LMM. The authors further concluded thatbnormalities of the LMM may explain referred leg pain inhe absence of other MRI abnormalities.

Hides et al [14] described a study of 26 subjects with unilat-ral LBP, some with referred lower extremity pain. The authorsssessed LMM atrophy among the subjects using ultrasoundmaging, noting marked asymmetry of the LMM that was local-zed to a single vertebral level in most cases. The authors notedhat the level of LMM asymmetry correlated with the neurologicevel of the symptoms in 24 of the 26 subjects in the study.

MAGING OF MULTIFIDUS ATROPHY

here are numerous publications describing the morpho-ogic assessment of the LMM using MRI, computed tomog-aphy (CT), and ultrasound (US).

Kader et al [13] used MRI in evaluating LMM atrophy,

hich they defined as muscular replacement with fat and n

brous tissues. The authors established a ranked gradingcale for LMM atrophy consisting of mild, moderate, andevere, corresponding to atrophy in less than 10% of CSA ofhe LMM, more than 10% and less than 50%, and more than0%, respectively (Figures 1-3). Two readers assessed theegree of LMM atrophy among the subjects, with good inter-bserver agreement noted.

Kjaer et al [12] described the blinded assessment of LMMtrophy in 854 MRI scans of adults and adolescents using arading scale similar to that used by Kader et al [13]. Theuthors found good intraobserver and interobserver agree-ent among the readers for the adult group. The authorsoted that fatty infiltration after LMM atrophy could beeasured in a noninvasive manner using MRI.Barker et al [15] performed MRI on 50 patients presenting

o a back pain clinic with unilateral persisting LBP. Theuthors measured the CSAs of the left and right psoas andMM and correlated their findings with the distribution anduration of symptoms among the subjects. A significantositive correlation between the side of LMM atrophy and theistribution of the LBP was found, as well as between theegree of atrophy and the duration of the symptoms.

CT has been evaluated as a relatively lower-cost and moreccessible alternative to MRI for assessment of LMM atrophy.anneels et al [16] described a study of CT assessment of 32 LBPatients and 23 active volunteers. The study design included CTf the CSA of the LMM at 3 levels. The authors found macro-copic atrophic changes in the LMM in 80% of the patients withBP, and that the atrophy in the LMM at the lowest levels of the

umbar spine correlated significantly with LBP status. The au-hors concluded that the most important finding from theirtudy was the finding of a significant correlation between LMMtrophy and radicular and nonradicular leg pain.

The utility and reliability of US relative to MRI scanning,ave also been evaluated as a means of assessing LMM CSA.ides et al [17] compared the CSA assessment of the LMM

mong healthy young adults using MRI and US, following atrict protocol The authors determined that LMM CSA coulde measured as accurately with US as with MRI, althoughhey did not assess the ability of US to discriminate the degreef atrophy in the LMM.

EFLEX INHIBITION MODEL OF LOCALIZEDULTIFIDUS ATROPHY

acintosh et al [11] described the morphology of the LMMy means of the dissection of 12 adult cadaver spines. Inontrast to previous determinations that each LMM is inner-ated by a number of spinal nerves, these authors found thathe LMM is divided into 5 distinct myotomes that are eachnnervated by a single spinal segment. They found that all

uscle fibers attaching to the spinous process or lamina of aarticular vertebra are segmentally innervated by the same

erve; the medial branch of the dorsal ramus that originates

icftsr

zaejiaListtlpjta

teoCratlm

amabbiiwanrn

occasmfbapl

atcmcdLaim

M

DmApbcntirscg

u

Fgf

144 Freeman et al ROLE OF THE LUMBAR MULTIFIDUS IN CHRONIC LOW BACK PAIN

nferior to the respective vertebra (Figure 4). Thus, the mus-le that moves a given spinal segment is supplied by the nerveor that segment. The authors postulated that from an ana-omic perspective, the shared innervation of the zygapophy-eal joints means that pain emanating from these joints couldesult in a reflex inhibition of the LMM at the same level.

Indahl et al [18] noted that the injection of saline into theygapophyseal joint in a porcine model resulted in decreasedctivity of the multifidus muscle. They concluded that theffect of the injection was to activate a stretch reflex in theoint capsule, which in turn excited inhibitory interneuronsn the spinal cord, which in turn inhibited the motor neuronsnd decreased the muscle response. The authors inferred thatMM atrophy seen in the human spine is a result of reflex

nhibition caused by afferent feedback from the zygapophy-eal joint, which in turn impedes the voluntary activation ofhe LMM. The authors pointed to their results as evidencehat multifidus atrophy seen in back pain patients is moreikely a result of dysfunction rather than disuse. The authorsostulated that inhibitory discharges from the zygapophyseal

oints may explain the efficacy of manual medical approacheso back pain such as manipulation and mobilization directedt the zygapophyseal joints.

Hodges et al [19] demonstrated, also in a porcine model,he rapid onset of LMM atrophy within 3 days after anxperimentally induced nerve root injury. After transectionf the medial branch of the L3 nerve root, the ipsilateral LMMSA adjacent to the L4, L5, and L6 spinous processes was

educed by 13%, 20%, and 12%, respectively, by 72 hoursfter injury. The changes were isolated to the side of injury;here were no differences in the CSA or muscular activityevels of the contralateral LMM, as determined by US assess-

ent. The authors also investigated the changes in the LMM

igure 1. Mild multifidus muscle atrophy, per Kader et al [13]rading criteria (less than 10% of CSA of muscle replaced with

iat).

fter experimentally induced disk injury in the porcineodel. A stab wound was introduced to the left anterolateral

spect of the L3-4 intervertebral disk, and this was followedy focal atrophy and a reduction of the CSA of the left LMMy 17%. The authors opined that it was unlikely that the

njury resulted from LMM denervation (owing to nerve rootnjury) as the atrophy was only at the level of the disk,hereas single nerve roots innervate three levels of LMM. The

uthors noted that the posttraumatic changes in the LMMoted in their study were undoubtedly attributable to aeduction in neural drive to the muscle, but that the mecha-ism that accounted for the reduction was unclear.

Hides et al [20] provided further evidence that the patternf LMM atrophy is local rather than general in a study thatompared multifidus size and bilateral symmetry betweenhronic LBP patients and healthy asymptomatic subjects. Theuthors reported that at the L4 and L5 levels, asymptomaticubjects were found to have significantly larger multifidususcles in comparison with chronic LBP patients. They also

ound that the greatest asymmetry was seen at the L5 verte-ral level in patients with unilateral pain presentations. Theuthors concluded that their findings support a clinical ap-roach to LBP that uses exercise therapy that focuses on

ocalized muscle impairments.Wallwork et al [21] demonstrated that LMM atrophy was

ssociated with a reduction in the ability to voluntarily con-ract the muscle. The authors used diagnostic US to measureontractions of the LMM by comparing the thickness of theuscle at rest to when it was contracted. Study subjects with

hronic LBP and LMM atrophy demonstrated significantlyecreased ability to perform isometric contractions of theirMM. These study results reinforced the findings of prioruthors and emphasized the clinical perspective that rehabil-tation in LBP patients may need to target localized impair-

ents in motor control.

ULTIFIDUS ATROPHY TREATMENT

anneels et al [4] assessed the efficacy of 3 different treat-ent modalities on multifidus CSA in chronic LBP patients.group of 59 patients were randomly allocated to one of 3

rograms: stabilization training, stabilization training com-ined with dynamic resistance, and stabilization trainingombined with dynamic-static resistance. Using CT scan-ing, LMM CSAs were measured before and after 10 weeks ofraining. The CSA of the LMM muscle was significantlyncreased at all vertebral levels only in the dynamic-staticesistance training group. The authors concluded that thetatic holding component between concentric and eccentricontraction phases was critical to induction of muscle re-rowth.

A randomized controlled trial performed by Van et al [22]sed real-time visual feedback by means of US to improve the

sometric contraction capability of the LMM in healthy sub-

jtgfwqmw

pis4Lf

ocatwrts

oapfilrsf

gshpiLdf

acwgospppdattaafm

C

Tt

F[o

Fi

145PM&R Vol. 2, Iss. 2, 2010

ects. The authors applied motor learning principles to trainhe subjects in voluntary contraction of the multifidus andave verbal feedback to one group, and verbal and visualeedback to the other. They found that providing the subjectsith visual feedback of their performance resulted in betteruality muscle contraction of the LMM as well as betteremory of how to contract the multifidi after reassessment aeek later.Sung [23] evaluated the question of whether an exercise

rogram designed to increase the ability to contract the LMMmproves functional status or reduces LBP. This author de-cribed a short-term cohort study that assessed the effect of a-week spinal stabilization exercise program in a group ofBP patients, reporting a significant improvement in LMMunction, as well as a reduction in LBP disability.

Hides et al [5] described the effects of stabilization trainingn LMM CSA among elite athletes (cricketers) with LBP,oncluding that LMM atrophy can exist in highly active, elitethletes. The authors demonstrated that the stabilizationraining increased the CSA of the LMM, and that the trainingas associated with a decrease in symptoms of LBP. Their

esults showed that specific training that aims at activation ofhe LMM is associated with both an increase in functionaltatus and a reduction in LBP.

In an earlier study by Hides et al [24], the authors dem-nstrated that LMM function recovery is not necessarilyssociated with the resolution of painful symptoms in LBPatients. The authors studied a group of 39 patients with arst episode of acute unilateral LBP and corresponding uni-

ateral segmental inhibition of the LMM. The patients wereandomly allocated to 2 groups, only one of which performedpecific exercises designed to reactivate the LMM through

igure 2. Moderate multifidus muscle atrophy, per Kader et al13] grading criteria (more than 10% but less than 50% of CSAf muscle replaced with fat).

acilitating an isometric contraction of these muscles. Thef

roup that did not perform the exercises had decreased LMMize at a 10-week follow-up examination even though theyad resumed normal levels of activity and had a remission ofainful symptoms, whereas the group that performed the

sometric contractions showed more rapid and completeMM recovery. The authors suggested that the continuedysfunction of the LMM in such patients may be one reasonor the high recurrence rates of LBP after an initial episode.

MacDonald et al [25] assessed muscular control of longnd short LMM fibers in 15 recurrent LBP patients and 19ontrol subjects. The authors used EMG to determinehether LMM control at the L5 level differed between theroups. They found that there were changes in motor controlf the LMM among the patients that primarily affected thehorter fibers, and that these changes were greater on thereviously painful side of the low back. These findings sup-ort the use of a specifically designed stabilization exerciserogram directed toward the LMM and specifically theeeper shorter fibers that serve to stabilize the spine. Theuthors concluded that the abnormal pattern of muscle con-rol they observed may leave the spine vulnerable to reinjury,hus predisposing the spine to recurrent episodes of LBP. Theuthors postulated that the observed motor control problemmong the recurrent LBP patients implied that pain andunctional performance should not be the only outcomeeasures after an acute episode of LBP.

ONCLUSIONS

he LMM are important stabilizers of the lumbar spine neu-ral zone, and atrophy of the muscle decreases the ability to

igure 3. Severe multifidus atrophy, per Kader et al [13] grad-ng criteria (more than 50% of CSA of muscle replaced with

at).

cTbmfatLcpcL

R

1

1

1

1

1

1

1

1

1

1

2

2

2

2

2

2

Tcgdn

Fv

146 Freeman et al ROLE OF THE LUMBAR MULTIFIDUS IN CHRONIC LOW BACK PAIN

ontrol the neutral zone and is strongly associated with LBP.his atrophy appears to help perpetuate an inhibitory feed-ack loop that begins with pain in the spine, possibly stem-ing from the intervertebral disks or zygapophyseal joints,

ollowed by reflex inhibition of the multifidus, and thentrophy and fatty replacement of the muscle. Amelioration ofhe LBP does not necessarily result in resumption of normalMM function, and decreased LMM function is likely impli-ated in recurrent LBP. Muscle training directed at teachingatients to activate their LMM is an important feature of anylinical approach to the patient with LBP who demonstratesMM dysfunction or atrophy.

EFERENCES1. Aure O, Nilsen J, Vasseljen O. Manual therapy and exercise therapy in

patients with chronic low back pain: A randomized, controlled trialwith 1-year follow-up. Spine 2003;28:525-532.

2. Suni J, Rinne M, Natri A, Pasanen M, Parkkari J, Alaranta H. Control ofthe lumbar neutral zone decreases low back pain and improves self-evaluated work ability. Spine 2006;31:E611-E620.

3. Von Korff M, Saunders K. The course of back pain in primary care.Spine 1996;21:2833-2837.

4. Danneels L, Vanderstraeten G, Cambier D, et al. Effects of three differ-ent training modalities on the cross sectional area of the lumbar multi-fidus muscle in patients with chronic low back pain. Br J Sports Med2001;35:186-191.

5. Hides J, Stanton W, McMahon S, Sims K, Richardson C. Effect ofstabilization training on multifidus muscle cross-sectional area amongyoung elite-cricketers with low back pain. J Orthop Sports Phys Ther2008;38:101-108.

6. Panjabi M. The stabilizing system of the spine. Part II. Neutral zone andinstability hypothesis. J Spinal Disord 1992;5:390-397.

7. Wilke H, Wolf S, Claes L, Arand M, Wiesend A. Stability of the lumbar

Transverse Process

Spinous Process

Intervertebral Disc

Facet Joint

Medial Branch ofThe Dorsal Ramus

Multifidus Muscle

igure 4. Schematic representation of orientation and inner-ation of a lumbar multifidus muscle.

spine with different muscle groups. Spine 1995;20:192-198.

8. Ward S, Kim C, Eng C, et al. Architectural analysis and intraoperativemeasurements demonstrate the unique design of the multifidus musclefor the lumbar spine stability. J Bone Joint Surg 2009;91:176-185.

9. Rosatelli AL, Ravichandiran K, Agur AM. Three-dimensional study ofthe musculotendinous architecture of lumbar multifidus and its func-tional implications. Clin Anat 2008;21:539-546.

0. MacDonald D, Moseley G, Hodges P. The lumbar multifidus muscle:does the evidence support the clinical beliefs. Man Ther 2006;11:254-263.

1. Macintosh J, Valencia F, Bogduk M, Munro R. The morphology of thelumbar multifidus. Clin Biomech 1986;1:196-204.

2. Kjaer P, Bendix T, Sorensen J, Korsholm L, Leboeuf-Yde C. Are MRI-defined fat infiltrations in the multifidus muscles associated with lowback pain? BMC Medicine 2007;5:2.

3. Kader D, Wardlaw D, Smith F. Correlation between the MRI changes inthe lumbar multifidus muscles and leg pain. Clin Radiol 2000;55:145-149.

4. Hides J, Stokes M, Saide M, Jull G, Cooper D. Evidence of lumbarmultifidus muscle wasting ipsilateral to symptoms in patients withacute/ subacute low back pain. Spine 1994;19:165-172.

5. Barker K, Shamley D, Jackson D. Changes in the cross-sectional area ofmultifidus and psoas in patients with unilateral back pain. Spine2004;29:E515-E519.

6. Danneels L, Vanderstraeten G, Cambier D, Witvrouw E, De Cuyper H.CT imaging of trunk muscles in chronic low back pain patients andhealthy control subjects. Eur Spine J 2000;9:266-272.

7. Hides J, Richardson C, Jull G. Magnetic resonance imaging and ultra-sonography of the lumbar multifidus muscle—comparison of twodifferent modalities. Spine 1995;20:54-58.

8. Indahl A, Kaigle A, Reikerås O, Holm S. Interaction between theporcine lumbar intervertebral disc, zygapophyseal joints, and paraspi-nal muscles. Spine 1997;22:2834-2840.

9. Hodges P, Holm A, Hansson T, Holm S. Rapid atrophy of the lumbarmultifidus follows experimental disc or nerve root injury. Spine 2006;31:2926-2933.

0. Hides J, Gilmore C, Stanton W, Bohlscheid E. Multifidus size andsymmetry among chronic LBP and healthy asymptomatic subjects. ManTher 2008;13:43-9. Epub 2006 Oct 27.

1. Wallwork TL, Stanton WR, Freke M, Hides JA. The effect of chroniclow back pain on size and contraction of the lumbar multifidus muscle.Man Ther 2009;14:496-500. Epub 2008 Nov 21.

2. Van K, Hides J, Richardson C. The use of real-time ultrasound imagingof lumbar multifidus muscle contraction in healthy subjects. J OrthopSports Ther 2006;36:920-925.

3. Sung P. Multifidus muscle median frequency before and after spinalstabilization exercises. Arch Phys Med Rehabil 2003;84:1313-1318.

4. Hides J, Richardson C, Jull G. Multifidus muscle recovery is notautomatic after resolution of acute, first-episode low back pain. Spine1996;21:2763-2769.

5. MacDonald D, Moseley G, Hodges P. Why do some patients keephurting their back? Evidence of ongoing back muscle dysfunctionduring remission from recurrent back pain. Pain 2009;142:183-188.

his CME activity is designated for 1.0 AMA PRA Category 1 Credit™ andan be completed online at me.aapmr.org. Log on to www.me.aapmr.org,o to Lifelong Learning (CME) and select Journal-based CME from therop down menu. This activity is FREE to AAPM&R members and $25 foron-members.