Motor Control Exercise for Persistent, Nonspecific Low Back Pain: A Systematic Review

Motor Control Exercise for Persistent,Nonspecific Low Back Pain:A Systematic ReviewLuciana G Macedo, Christopher G Maher, Jane Latimer, James H McAuley

Background. Previous systematic reviews have concluded that the effectivenessof motor control exercise for persistent low back pain has not been clearlyestablished.

Objective. The objective of this study was to systematically review randomizedcontrolled trials evaluating the effectiveness of motor control exercises for persistentlow back pain.

Methods. Electronic databases were searched to June 2008. Pain, disability, andquality-of-life outcomes were extracted and converted to a common 0 to 100 scale.Where possible, trials were pooled using Revman 4.2.

Results. Fourteen trials were included. Seven trials compared motor control ex-ercise with minimal intervention or evaluated it as a supplement to another treat-ment. Four trials compared motor control exercise with manual therapy. Five trialscompared motor control exercise with another form of exercise. One trial comparedmotor control exercise with lumbar fusion surgery. The pooling revealed that motorcontrol exercise was better than minimal intervention in reducing pain at short-termfollow-up (weighted mean difference��14.3 points, 95% confidence interval[CI]��20.4 to �8.1), at intermediate follow-up (weighted mean difference��13.6points, 95% CI��22.4 to �4.1), and at long-term follow-up (weighted mean differ-ence��14.4 points, 95% CI��23.1 to �5.7) and in reducing disability at long-termfollow-up (weighted mean difference��10.8 points, 95% CI��18.7 to �2.8). Motorcontrol exercise was better than manual therapy for pain (weighted mean differ-ence��5.7 points, 95% CI��10.7 to �0.8), disability (weighted mean differ-ence��4.0 points, 95% CI��7.6 to �0.4), and quality-of-life outcomes (weightedmean difference��6.0 points, 95% CI��11.2 to �0.8) at intermediate follow-upand better than other forms of exercise in reducing disability at short-term follow-up(weighted mean difference��5.1 points, 95% CI��8.7 to �1.4).

Conclusions. Motor control exercise is superior to minimal intervention andconfers benefit when added to another therapy for pain at all time points and fordisability at long-term follow-up. Motor control exercise is not more effective thanmanual therapy or other forms of exercise.

LG Macedo, PT, MSc, is a PhD stu-dent at The George Institute forInternational Health, The Univer-sity of Sydney, PO Box M201, Mis-senden Rd, Camperdown, Syd-ney, New South Wales, 2050Australia. Address all correspon-dence to Ms Macedo at:[email protected].

CG Maher, PT, PhD, is Director,Musculoskeletal Division, TheGeorge Institute for InternationalHealth, The University of Sydney.

J Latimer, PT, PhD, is AssociateProfessor, The George Institute forInternational Health, The Univer-sity of Sydney.

JH McAuley, PhD, is ResearchManager, The George Institute forInternational Health.

[Macedo LG, Maher CG, Latimer J,McAuley JH. Motor control exer-cise for persistent, nonspecific lowback pain: a systematic review.Phys Ther. 2009;89:9–25.]

© 2009 American Physical TherapyAssociation

Research Report

Post a Rapid Response orfind The Bottom Line:www.ptjournal.org

January 2009 Volume 89 Number 1 Physical Therapy f 9

Low back pain (LBP) is one of themain causes of disability, and,despite its high prevalence, the

source of pain is not established inthe majority of cases and the term“nonspecific low back pain” isused.1–4 One factor that has beenproposed as important in the genesisand persistence of nonspecific LBP isstability and control of the spine.4

Studies of individuals with LBP haveidentified impairments in the controlof the deep trunk muscles (eg, trans-versus abdominis and multifidus) re-sponsible for maintaining the stabil-ity of the spine.5–8 For example,activity of the transversus abdominismuscles9 and the multifidus muscles7

is delayed during arm movements(that challenge the stability of thespine) in individuals with LBP. Fur-thermore, there is evidence of de-creased cross-sectional area10 andincreased fatiguability11 and a sug-gestion of increased intramuscularfat in the paraspinal muscles of indi-viduals with LBP.12 Therefore, theo-retically, an intervention that aims tocorrect the changes occurring in thedeep trunk muscles and that targetsthe restoration of control and coor-dination of these muscles should beeffective in the management of per-sistent LBP.

Motor control exercise was devel-oped based on the principle that in-dividuals with LBP have a lack ofcontrol of the trunk muscles. Theidea is to use a motor learning ap-proach to retrain the optimal controland coordination of the spine. Theintervention involves the training ofpreactivation of the deep trunk mus-cles, with progression toward morecomplex static, dynamic, and func-tional tasks integrating the activationof deep and global trunk muscles.13,14

Although a number of laboratorystudies supporting the underlyingmechanism of action of motor con-trol exercises have been published inthe last decades,5,9,15 the clinical ef-

fectiveness of motor control exer-cise for persistent LBP is still un-clear.5,9,15 Three systematic reviewsof motor control exercise have beenpublished16–18; however, the authorsof these reviews searched the litera-ture only up until October 2005.Hauggaard and Persson,17 the au-thors of the latest published review,included 10 trials testing the efficacyof motor control for acute, subacute,and chronic LBP. The review used asimple descriptive approach to sum-marize the results of each individualtrial. Rackwitz et al18 summarizedthe results of 7 randomized con-trolled trials of acute, subacute, andchronic LBP, and although they useda better approach to summarize theavailable evidence, no meta-analytical analysis with pooling ofthe data was used. Ferreira et al16

summarized the results of 13 ran-domized controlled trials of recur-rent, acute, subacute, and chronicLBP and cervical pain. This reviewwas the only one that included ameta-analytical approach; however,only a few trials were pooled, limit-ing the generalization of the re-sults. A meta-analytical approach issuperior to the other forms of analy-sis for systematic reviews because itprovides a treatment effect size with95% confidence interval (CI).

Consistent with the Cochrane Col-laboration,19 we felt that an updatedreview incorporating new random-ized controlled trials would make auseful contribution to the literature.In addition, a meta-analytical ap-proach, which has not been widelyused in the previous published sys-tematic reviews, can potentially adduseful information about the magni-tude of the effect of motor controlexercises. Because our main interestwas to study persistent LBP andguidelines suggest that persistentand acute LBP should be consideredseparately,19–21 we included only tri-als studying patients with LBP thatpersisted beyond the acute phase.

The term “persistent low back pain”is used to describe subacute,chronic, and recurrent pain. Thus,the objective of this study was tosystematically review randomizedcontrolled trials testing the effect ofmotor control exercise in patientswith persistent, nonspecific LBP.

MethodData Sources and SearchesA computerized electronic searchwas performed to identify relevantarticles. The search was conductedon MEDLINE (1950 to June 2008),CINAHL (1982 to June 2008), AMED(1985 to June 2008), PEDro (to June2008), and EMBASE (1988 to June2008). Key words relating to the do-mains of randomized controlled tri-als and back pain were used, as rec-ommended by the Cochrane BackReview Group.19 Terms for motorcontrol and specific stabilizationexercises were extracted from thereview by Ferreira et al.16 Subjectsubheadings and word truncations,according to each database, wereused. There was no languagerestriction.

One reviewer (LGM) screenedsearch results for potentially eligiblestudies, and 2 reviewers (LGM,CGM) independently reviewed arti-cles for eligibility. A third indepen-dent reviewer (JL) resolved any dis-agreement about inclusion of trials.Authors were contacted if more in-formation about the trial was neededto allow inclusion of the study. Re-searchers who published in the areawere contacted to help identify grayliterature and articles in press. Cita-tion tracking was performed usingISI Web of Science, and a manualsearch of the reference lists of previ-ous reviews and the eligible trialswas performed.

Study SelectionThe reviewers followed a researchprotocol, developed prior to the be-ginning of the review, that included

Motor Control Exercise for Persistent, Nonspecific LBP

10 f Physical Therapy Volume 89 Number 1 January 2009

a checklist for inclusion criteria. Ar-ticles were eligible for inclusionif they were randomized or quasi-randomized controlled trials compar-ing motor control exercise with aplacebo treatment, no treatment, oranother active treatment or whenmotor control exercise was added asa supplement to other interventions.When motor control exercise wasused in addition to other treatments,motor control exercises had to rep-resent at least 40% of the total treat-ment program. This criterion wasjudged by reading the description ofthe treatment with the reviewermaking a global yes/no judgment.

Trials were considered to have eval-uated motor control exercise if theexercise treatment was described asmotor control or specific spinal sta-bilization or core stability exerciseand where the protocol describedexercise targeting specific trunkmuscles in order to improve controland coordination of the spine andpelvis.

Randomized or quasi-randomizedcontrolled trials were included ifthey explicitly reported that a crite-rion for entry was nonspecific LBP(with or without leg pain) of at least6 weeks’ duration (nonacute LBP) orrecurrent LBP. Studies evaluating in-dividuals of all age groups of eithersex were included. Trials were in-cluded if one of the following out-come measures had been reported:pain, disability, quality of life, returnto work, or recurrence.

Data Extraction and QualityAssessmentThe methodological quality of the tri-als was assessed using the PEDroscale,22 with scores extracted fromthe PEDro database. Assessment ofquality of trials in the PEDro databasewas performed by 2 trained inde-pendent raters, and disagreementswere resolved by a third rater.23 Onestudy24 was extracted from a confer-

ence proceeding, and, therefore, thePEDro score was not available inthe database. However, 2 PEDro rat-ers evaluated the information avail-able in the abstract and in an initialversion of a manuscript, and a PEDroscore was given. Methodologicalquality was not an inclusioncriterion.

Three independent reviewers (LGM,CGM, JL) extracted data from eachincluded study using a standardizedextraction form. Mean scores, stan-dard deviations, and sample sizeswere extracted from the studies.When this information was not pro-vided in the trial, the values werecalculated or estimated using meth-ods recommended in the CochraneHandbook for Systematic Reviewsof Interventions.25 When there wasinsufficient information about out-comes to allow data analysis, the au-thors of the study were contacted,and all authors replied to ourinquiries.24,26–28

Outcomes were extracted for painand disability for short-termfollow-up (less than 3 monthsafter randomization), intermediatefollow-up (at least 3 months but lessthan 12 months after randomiza-tion), and long-term follow-up (12months or more after randomiza-tion). When there were multipletime points that fell within the samecategory, the one that was closer tothe end of the treatment for theshort-term follow-up, closer to 6months for the intermediate follow-up, and closer to 12 months for thelong-term follow-up was used. Thesereferences for time points werebased on guidelines from the Coch-rane Back Review Group. Scores forpain and disability were converted toa 0 to 100 scale.29

Data Synthesis and AnalysisThe studies were grouped into 4treatment contrasts: (1) motor con-trol versus minimal intervention (no

intervention, general practitionercare, education) or motor control asa supplement, (2) motor control ver-sus spinal manipulative therapy, (3)motor control versus exercise, and(4) motor control versus surgery(lumbar fusion). Results were pooledwhen trials were considered suffi-ciently homogenous with respect toparticipant characteristics, interven-tions, and outcomes. I2 was calcu-lated using RevMan 4.2* to analyzestatistical heterogeneity. I2 describesthe percentage of the variability ineffect estimates that is due to heter-ogeneity rather than sampling error(chance). A value greater than 50%may be considered substantial heter-ogeneity.25 When trials were statisti-cally homogeneous (I2�50%),pooled effects (weighted mean dif-ference) were calculated using afixed-effect model. When trials werestatistically heterogeneous (I2�50%)pooled estimates of effect (weightedmean difference) were obtained us-ing a random-effects model.25 Whenthere was a single trial for the com-parison, results were expressed asmean differences and 95% CI.

ResultsStudy SelectionThe initial electronic database searchresulted in a total of 1,052 articles.Of these, 42 were selected as poten-tially eligible based on their title andabstract. Through a Web of Sciencesearch of these articles, 3 other po-tentially eligible articles were identi-fied. A total of 45 potentially eligiblearticles were considered for inclu-sion, with only 14 eligible for inclu-sion in this review (Fig. 1). Reasonsfor exclusion are shown in Figure 1for those articles2,3,15,30–57 that wereexcluded from this review. Only 1 ofthe 26 experts contacted sent infor-mation to us on a new trial forinclusion.

* Copenhagen, Denmark: The Nordic Coch-rane Centre, The Cochrane Collaboration,2003.



A number of randomized controlledtrials that were included in previoussystematic reviews of motor controlexercises were not included in thisreview. Reasons for exclusion in-cluded: patients had acute but notpersistent back pain,15,51,53 patientshad neck pain and headache but notback pain,58 the trial did not use amotor control intervention accord-ing to our review definition,56 and

the trial did not have the outcomesof interest.59,60 Four new tri-als13,24,26,61 that were not included inany of the previously published re-views were included in this review,accounting for the addition of 560patients.

Methodological QualityThe methodological quality assess-ment using the PEDro scale revealed

a mean score of 6 (range�2–8).Blinding of the therapist and blind-ing of the subject were not used inany of the trials, as would be ex-pected for an exercise therapy study.An intention-to-treat analysis wasused in 36% of the trials, and alloca-tion concealment was present in58% of the trials. One of the articles24

included in the review was from aconference proceeding, and, there-

Figure 1.Flow chart of systematic review inclusion and exclusion. RCT�randomized controlled trial.



fore, not much information on theconduct of the trial was available.With the limited information avail-able, this trial received a score of 2on the PEDro scale and was the onlytrial that was a quasi-randomizedcontrolled trial.24

Study CharacteristicsThe 14 randomized controlled trialsincluded in this review comparedmotor control exercise against an-other treatment or against no treat-ment (Tabs. 1 and 2). No placebo-controlled trials were identified.Trials were grouped into 4 treatmentcontrasts: (1) motor control exerciseversus minimal intervention or mo-tor control exercise as a supplement,(2) motor control exercise versusmanual therapy, (3) motor controlexercise versus other forms of exer-cise, and (4) motor control exerciseversus surgery.

Seven trials (603 patients) were in-cluded in the first treatment con-trast: 4 trials (343 patients) that com-pared motor control exercise withminimal intervention (no interven-tion, general practitioner care, or ed-ucation)14,27,62,63 and 3 trials (260 pa-tients) that used motor controlexercise as a supplement to othertreatment (general exercise or usualphysical therapy.28,64,65 Four trials(523 patients) compared motor con-trol exercise with manual therapy(high- or low-velocity trust).13,26,64,66

Five trials (508 patients) comparedmotor control exercise with anotherform of exercise therapy (pain man-agement, general exercises, or theMcKenzie approach).13,24,26,61,67 Onetrial (61 patients) compared motorcontrol exercise with lumbar fusionsurgery.68 The characteristics of themotor control exercise programsthat were evaluated in each trial areprovided in Table 2.

Motor Control Exercise VersusMinimal Intervention or MotorControl Exercise as a SupplementOf the 7 studies included in thistreatment contrast, 4 compared mo-tor control exercise with a minimalintervention program (usual generalpractitioner care or no interven-tion)14,27,62,63 and 3 compared motorcontrol exercise as a supplement toanother intervention versus thisother intervention alone.28,64,65

Methodological quality of the articlesranged from 4 to 8. Data for pain,disability, and quality of life wereavailable for pooling at short-term,intermediate, and long-term follow-up. Data were pooled using arandom-effects model for all compar-isons except for quality of life at in-termediate and long-term follow-ups,where a fixed-effects model wasused because I2 was smaller than50%.

The pooled results favored motorcontrol exercise for pain and disabil-ity outcomes at each follow-up, with4 of the 6 estimates of treatmenteffect being statistically significant.The random-effects model showed astatistically significant decrease inpain favoring motor control exerciseat short-term follow-up (weightedmean difference [on a 0–100scale]��14.3 points, 95%CI��20.4 to �8.1), intermediatefollow-up (weighted mean differ-ence�13.6 points, 95% CI��22.4 to�4.1), and long-term follow-up(weighted mean difference��14.4points, 95% CI��23.1 to �5.7) andin reducing disability at long-termfollow-up (weighted mean differ-ence��10.8 points, 95% CI��18.7to �2.8) (Fig. 2). There was no evi-dence that motor control exercisewas effective for improving qualityof life.

Motor Control Exercise VersusManual TherapyFour trials13,26,64,66 compared motorcontrol exercise with manual ther-

apy, with pain and disability out-comes measured at short-term, inter-mediate, and long-term follow-upsand quality of life measured at inter-mediate and long-term follow-ups.The methodological quality of the ar-ticles ranged from 4 to 8. Because I2

was smaller than 50% for all timepoints, a fixed-effects model wasused to pool the results. The pooledeffects for pain and disability out-comes favored motor control exer-cise, but the effects were alwayssmall and reached statistical signifi-cance for only 2 of the 6 estimates.There was a significant difference be-tween treatment groups favoringmotor control exercise for pain anddisability at intermediate follow-up(weighted mean difference��5.7points, 95% CI��10.7 to �0.8 forpain and weighted mean differ-ence��4.0 points, 95% CI��7.6 to�0.4 for disability) (Fig. 3). Thepooled estimates of treatment effectson quality of life were small, favoringmotor control exercise at short-termfollow-up and favoring manual ther-apy at long-term follow-up.

Motor Control Exercise VersusOther Forms of ExerciseFive trials13,24,26,61,67 compared mo-tor control exercise with anotherform of exercise therapy. The meth-odological quality of the trials rangedfrom 2 to 8. The trial with a method-ological quality score of 2 had itsPEDro score assessed from a confer-ence proceeding and some informa-tion given by the authors.24 Resultswere pooled for pain and disabilityat short-term, intermediate, and long-term follow-ups. Because I2 wasgreater than 50% for pain at short-term follow-up and for disability atlong-term follow-up, pooled effectsfor these time points were calculatedusing a random-effects model. Allother pooled effects were calculatedusing a fixed-effects model. All esti-mates of treatment effect were small.Five of the 6 estimates favored motorcontrol exercise; however, only one



Tab

le1.

Det

ails

ofth

eIn

clud

edRa

ndom

ized

Con

trol

led

Tria

lsa

Art

icle

Pat

ien

tC

har

acte

rist

ics,

Sam

ple

Size

,an

dD

ura

tio

no

fC

om

pla

int

Inte

rven

tio

ns

Ou

tco

mes

(Mea

sure

)P

EDro

Sco

reA

rtic

leIn

clu

ded

inP

revi

ou

sR

evie

ws

Mo

tor

con

tro

lex

erci

ses

vers

us

min

imal

inte

rven

tio

no

rm

oto

rco

ntr

ol

exer

cise

sas

asu

pp

lem

ent

Nie

mis

toet

al,6

2

2003

Patie

nts

recr

uite

dfr

omad

vert

isem

ent

Age

d24

–46

yM

ain

excl

usio

ncr

iterio

n:ne

urol

ogic

alsi

gns

orp

rior

back

surg

ery

N�

204

Dur

atio

nof

LBP

�3

mo

Mot

orco

ntro

lexe

rcis

es�

mus

cle

ener

gyvs

usua

lge

nera

lpra

ctiti

oner

care

(edu

catio

n)

Pain

(VA

S)D

isab

ility

(OD

I)Q

ualit

yof

life

(hea

lth-r

elat

edq

ualit

yof

life)

8In

clud

edin

Ferr

eira

etal

,16

2006

;Ra

ckw

itzet

al,1

820

06;

and

Hau

ggaa

rdet

al,1

720

07

Koum

anta

kis

etal

,65

2005

Patie

nts

from

anor

thop

edic

clin

icin

aho

spita

land

gene

ral

pra

ctiti

oner

sM

ain

excl

usio

ncr

iterio

n:p

rior

back

surg

ery

orra

diol

ogic

alsi

gns

ofsp

inal

inst

abili

tyN

�55

Dur

atio

nof

LBP

�6

wk

Mot

orco

ntro

lexe

rcis

es�

gene

rale

xerc

ises

vsge

nera

lex

erci

ses

only

Pain

(VA

S)D

isab

ility

(RM

-24)

7In

clud

edin

Ferr

eira

etal

,16

2006

;an

dH

augg

aard

etal

,17

2007

O’S

ulliv

anet

al,1

4

1997

Patie

nts

with

spon

dylo

lysi

sor

spon

dylo

listh

esis

Age

d16

–49

yM

ain

excl

usio

ncr

iterio

n:ne

urol

ogic

alsi

gns

orin

flam

mat

ory

join

tdi

seas

eN

�42

Dur

atio

nof

LBP

�3

mon

ths

Mot

orco

ntro

lexe

rcis

esvs

usua

lgen

eral

pra

ctiti

oner

care

Pain

(sho

rt-f

orm

McG

illVA

S)D

isab

ility

(OD

I)7

Incl

uded

inFe

rrei

raet

al,1

6

2006

;Ra

ckw

itzet

al,1

820

06;

and

Hau

ggaa

rdet

al,1

720

07

Stug

eet

al,2

820

04Pa

tient

sfr

omhe

alth

care

pra

ctiti

oner

sPe

lvic

gird

lep

ain

late

ralt

oL5

–S1

Mai

nex

clus

ion

crite

rion:

neur

olog

ical

sign

sN

�81

Dur

atio

nof

LBP

�6

wk

Mot

orco

ntro

lexe

rcis

es�

usua

lphy

sica

lthe

rap

yvs

usua

lphy

sica

lthe

rap

yon

ly

Pain

(VA

Sp

ain

even

ing)

Dis

abili

ty(O

DI)

7In

clud

edin

Ferr

eira

etal

,16

2006

Mos

eley

,27

2002

Patie

nts

from

gene

ralp

ract

ition

ers

and

phy

sica

lthe

rap

ycl

inic

sM

ain

excl

usio

ncr

iterio

n:w

orse

ning

neur

olog

ical

sign

sN

�57

Dur

atio

nof

LBP

�2

mo

Mot

orco

ntro

lexe

rcis

es�

man

ualt

hera

py

�ed

ucat

ion

vsus

ualg

ener

alp

ract

ition

erca

re

Pain

(bac

kp

ain

NRS

0–10

)D

isab

ility

(RM

-18)

6In

clud

edin

Ferr

eira

etal

,16

2006

;an

dRa

ckw

itzet

al,1

8

2006

Shau

ghne

ssy

etal

,63

2004

Patie

nts

from

orth

oped

iccl

inic

sA

ged

20–6

0y

Mai

nex

clus

ion

crite

rion:

neur

olog

ical

sign

sor

infla

mm

ator

yjo

int

dise

ase

N�

41D

urat

ion

ofLB

P�

3m

o

Mot

orco

ntro

lexe

rcis

esvs

noin

terv

entio

nPa

in(S

F-36

bodi

lyp

ain)

Dis

abili

ty(R

M-2

4)Q

ualit

yof

life

(SF-

36ge

nera

lhe

alth

)

5In

clud

edin

Hau

ggaa

rdet

al,1

7

2007

Gol

dby

etal

,64

2006

19

Patie

nts

from

phy

sica

lthe

rap

yde

par

tmen

tof

aho

spita

lA

ged

18–6

5y

Mai

nex

clus

ion

crite

rion:

neur

olog

ical

sign

sor

prio

rba

cksu

rger

yN

�12

4D

urat

ion

ofLB

P�

3w

k

Mot

orco

ntro

lexe

rcis

es�

educ

atio

nvs

educ

atio

non

lyPa

in(b

ack

pai

nN

RS0–

100)

Dis

abili

ty(O

DI)

Qua

lity

oflif

e(N

othi

ngha

mH

ealth

Profi

le)

4In

clud

edin

Ferr

eira

etal

,16

2006

;an

dH

augg

aard

etal

,17

2007

(Con

tinue

d)



Tab

le1.

Con

tinue

d

Art

icle

Pat

ien

tC

har

acte

rist

ics,

Sam

ple

Size

,an

dD

ura

tio

no

fC

om

pla

int

Inte

rven

tio

ns

Ou

tco

mes

(Mea

sure

)P

EDro

Sco

reA

rtic

leIn

clu

ded

inP

revi

ou

sR

evie

ws

Mo

tor

con

tro

lex

erci

ses

vers

us

man

ual

ther

apy

Ferr

eira

etal

,13

2007

Patie

nts

seek

ing

care

from

phy

sica

lthe

rap

yde

par

tmen

tsof

pub

licho

spita

lsA

ged

18–8

0y

Mai

nex

clus

ion

crite

rion:

neur

olog

ical

sign

sor

prio

rba

cksu

rger

yN

�16

0D

urat

ion

ofLB

P�

3m

o

Mot

orco

ntro

lexe

rcis

esvs

spin

alm

anip

ulat

ive

ther

apy

Pain

(VA

S)D

isab

ility

(RM

-24)

8N

otin

clud

edin

pre

viou

sre

view

s

Crit

chle

yet

al,2

6

2007

Patie

nts

recr

uite

dfr

omre

ferr

als

bysp

ecia

lists

orp

rimar

yca

rep

ract

ition

ers

top

hysi

calt

hera

py

dep

artm

ents

ofho

spita

lsA

ged

18y

orol

der

With

orw

ithou

tle

gsy

mp

tom

sor

neur

olog

icsi

gns

Mai

nex

clus

ion

crite

rion:

prio

rsp

inal

surg

ery,

hem

atol

ogic

dise

ase,

orha

dp

hysi

calt

hera

py

inth

ela

st6

mo

N�

143

Dur

atio

nof

LBP

�12

wk

Mot

orco

ntro

lexe

rcis

esvs

man

ualt

hera

py

�ho

me

exer

cise

svs

pai

nm

anag

emen

tp

rogr

am

Pain

(VA

S)D

isab

ility

(RM

-24)

Qua

lity

oflif

e(E

Q-5

D)

7N

otin

clud

edin

pre

viou

sre

view

s

Rasm

usse

n-Ba

rret

al,6

620

03N

�47

Dur

atio

nof

LBP

�6

wk

Mot

orco

ntro

lexe

rcis

esvs

spin

alm

anip

ulat

ive

ther

apy

Pain

(VA

S)D

isab

ility

(OD

I)5

Incl

uded

inFe

rrei

raet

al,1

6

2006

;an

dRa

ckw

itzet

al,1

8

2006

Gol

dby

etal

,64

2006

Patie

nts

from

phy

sica

lthe

rap

yde

par

tmen

tof

aho

spita

lA

ged

18–6

5y

Mai

nex

clus

ion

crite

rion:

neur

olog

ical

sign

sor

prio

rba

cksu

rger

yN

�17

3D

urat

ion

ofLB

P�

3w

k

Mot

orco

ntro

lexe

rcis

es�

educ

atio

nvs

spin

alm

anip

ulat

ive

ther

apy

�ed

ucat

ion

Pain

(bac

kp

ain

NRS

0–10

0)D

isab

ility

(OD

I)Q

ualit

yof

life

(Not

hing

ham

Hea

lthPr

ofile

)

4In

clud

edin

Ferr

eira

etal

,16

2006

;an

dH

augg

aard

etal

,17

2007

Mo

tor

con

tro

lex

erci

ses

vers

us

oth

erfo

rms

of

exer

cise

Ferr

eira

etal

,13

2007

Patie

nts

seek

ing

care

from

phy

sica

lthe

rap

yde

par

tmen

tsof

pub

licho

spita

lsA

ged

18–8

0y

Mai

nex

clus

ion

crite

rion:

neur

olog

ical

sign

sor

prio

rba

cksu

rger

yN

�16

0D

urat

ion

ofLB

P�

3m

o

Mot

orco

ntro

lexe

rcis

esvs

gene

rale

xerc

ises

Pain

(VA

S)D

isab

ility

(RM

-24)

8N

otin

clud

edin

pre

viou

sre

view

s

Crit

chle

yet

al,2

6

2007

Patie

nts

recr

uite

dfr

omre

ferr

als

bysp

ecia

lists

orp

rimar

yca

rep

ract

ition

ers

top

hysi

calt

hera

py

dep

artm

ents

ofho

spita

lsA

ged

18y

orol

der

With

orw

ithou

tle

gsy

mp

tom

sor

neur

olog

icsi

gns

Mai

nex

clus

ion

crite

rion:

prio

rsp

inal

surg

ery,

hem

atol

ogic

aldi

seas

e,or

had

phy

sica

lthe

rap

yin

the

last

6m

oN

�14

1D

urat

ion

ofLB

P�

12w

k

Mot

orco

ntro

lexe

rcis

esvs

man

ualt

hera

py

�ho

me

exer

cise

svs

pai

nm

anag

emen

tp

rogr

am

Pain

(VA

S)D

isab

ility

(RM

-24)

Qua

lity

oflif

e(E

Q-5

D)

7N

otin

clud

edin

pre

viou

sre

view

s

(Con

tinue

d)



Tab

le1.

Con

tinue

d

Art

icle

Pat

ien

tC

har

acte

rist

ics,

Sam

ple

Size

,an

dD

ura

tio

no

fC

om

pla

int

Inte

rven

tio

ns

Ou

tco

mes

(Mea

sure

)P

EDro

Sco

reA

rtic

leIn

clu

ded

inP

revi

ou

sR

evie

ws

Klad

nyet

al,6

720

03Pa

tient

sse

ntto

the

outp

atie

ntre

habi

litat

ion

dep

artm

ent

due

toba

ckp

ain

Age

d18

–55

yPa

tient

sw

ithor

with

out

radi

atio

nor

with

orw

ithou

tdi

skhe

rnia

orp

rotr

usio

nM

ain

excl

usio

ncr

iteria

:p

rior

spin

alsu

rger

y,ar

thrit

isof

the

join

ts,

inju

ries,

ortr

aum

aN

�99

Suba

cute

and

chro

nic

Mot

orco

ntro

lexe

rcis

es�

gene

rale

xerc

ises

vsge

nera

lex

erci

ses

�m

anua

lthe

rap

y

Pain

(bac

kp

ain

NRS

)D

isab

ility

(OD

I)5

Incl

uded

inFe

rrei

raet

al,1

6

2006

;Ra

ckw

itzet

al,1

820

06;

and

Hau

ggaa

rdet

al,1

720

07

Mill

eret

al,6

120

05Pa

tient

sfr

oman

outp

atie

ntp

hysi

calt

hera

py

clin

icA

ged

abov

e18

yM

ain

excl

usio

ncr

iterio

n:m

ore

than

one

back

surg

ery

orsy

stem

icin

flam

mat

ory

dise

ase

N�

30D

urat

ion

ofLB

P�

7w

k

Mot

orco

ntro

lexe

rcis

esvs

McK

enzi

eap

pro

ach

Pain

(VA

S)D

isab

ility

(fun

ctio

nals

tatu

s0–

100)

5N

otin

clud

edin

pre

viou

sre

view

s

Stev

ens

etal

,24

2007

Patie

nts

with

nons

pec

ific

LBP

from

the

phy

sica

lmed

icin

ean

dor

thop

edic

surg

ery

dep

artm

ent

ofa

hosp

ital

Age

d18

–65

yM

ain

excl

usio

ncr

iteria

:sp

ecifi

cLB

P,ra

dicu

lar

sym

pto

ms,

back

surg

ery,

and

neur

olog

icor

syst

emic

cond

ition

N�

78D

urat

ion

ofLB

P�

3m

oor

recu

rren

t

Mot

orco

ntro

lexe

rcis

es�

man

ualt

hera

py

(10%

)vs

gene

rale

xerc

ises

oftr

unk

mus

cle

func

tion

and

coor

dina

tion

Pain

(VA

S)D

isab

ility

(QBP

DS)

Qua

lity

oflif

e(S

F-36

gene

ral

heal

th)

2N

otin

clud

edin

pre

viou

sre

view

s

Mo

tor

con

tro

lex

erci

ses

vers

us

surg

ery

Brox

etal

,68

2003

Patie

nts

from

dep

artm

ents

ofor

thop

edic

surg

ery,

neur

osur

gery

,p

hysi

calm

edic

ine,

and

reha

bilit

atio

nA

ged

25–6

0y

Spin

ede

gene

ratio

nor

spon

dylo

sis

had

tobe

pre

sent

Mai

nex

clus

ion

crite

rion:

neur

olog

ical

sign

sor

prio

rba

cksu

rger

yN

�61

Dur

atio

nof

LBP

�1

y

Mot

orco

ntro

lexe

rcis

es�

cogn

itive

beha

vior

alth

erap

yvs

surg

ery

Pain

(bac

kp

ain

0–10

0sc

ale)

Dis

abili

ty(O

DI)

Qua

lity

oflif

e(li

fesa

tisfa

ctio

nsc

ale)

8In

clud

edin

Ferr

eira

etal

,16

2006

aLB

P�lo

wba

ckp

ain,

OD

I�O

swes

try

Dis

abili

tyIn

dex,

VAS�

visu

alan

alog

scal

e,RM

-18�

18-it

emRo

land

-Mor

risD

isab

ility

Que

stio

nnai

re,

RM-2

4�24

-item

Rola

nd-M

orris

Dis

abili

tyQ

uest

ionn

aire

,N

RS�

num

eric

alra

ting

scal

e,SF

-36�

Med

ical

Out

com

eSt

udy

36-I

tem

Shor

t-Fo

rmH

ealth

Surv

ey,

QBP

DS�

Que

bec

Back

Pain

Dis

abili

tySc

ale,

EQ-5

D�

Euro

Qol

que

stio

nnai

re.



Tab

le2.

Det

ails

ofth

eM

otor

Con

trol

Exer

cise

s

Art

icle

Du

rati

on

of

Mo

tor

Co

ntr

ol

Inte

rven

tio

nP

rog

ress

ion

Ru

leH

om

eP

rog

ram

Ad

her

ence

Mea

n(S

D)

Feed

bac

k

Brox

etal

,68

2003

5-w

kin

terv

entio

n(1

sess

ion

inth

efir

stw

eek,

2w

kof

hom

ep

rogr

am,

and

anot

her

2w

kof

trea

tmen

t)A

vera

gedu

ratio

nw

asab

out

25h

per

wee

k

Not

stat

ed2

wk

ofho

me

pro

gram

Adh

eren

cew

as3

(7)

sess

ions

per

pat

ient

Not

stat

ed

Crit

chle

yet

al,2

6

2007

8se

ssio

nsof

90m

inPr

ogre

ssio

nw

asba

sed

onth

eab

ility

ofth

ep

atie

nts

tom

aint

ain

ast

able

and

min

imal

lyp

ainf

ulsp

ine.

The

exer

cise

sai

med

toim

pro

vem

uscl

em

otor

cont

rolt

op

rovi

dedy

nam

icse

gmen

tals

tabi

lity

for

the

lum

bar

spin

e.

Not

stat

edN

otst

ated

Not

stat

ed

Ferr

eira

etal

,13

2007

12se

ssio

nsin

8w

kPr

ogre

ssio

nby

inco

rpor

atin

gm

ore

func

tiona

lp

ositi

ons

and

trai

ning

the

coor

dina

tion

ofal

ltr

unk

mus

cles

durin

gth

ose

func

tiona

ltas

ks

Not

stat

edA

dher

ence

was

9.2

(3.4

)se

ssio

nsp

erp

atie

ntRe

al-t

ime

ultr

asou

nd

Gol

dby

etal

,64

2006

1se

ssio

nof

11⁄2

hp

erw

eek

for

10w

kN

otst

ated

Not

stat

edN

otst

ated

Not

stat

ed

Klad

nyet

al,6

720

03N

otst

ated

Not

stat

edN

otst

ated

Stat

edon

lyth

atp

atie

nts

did

16.4

(4.8

)d

ofm

otor

cont

rol�

9.5

(3.4

)d

ofge

nera

lex

erci

ses

Real

-tim

eul

tras

ound

Koum

anta

kis

etal

,65

2005

2se

ssio

nsof

30to

45m

inp

erw

eek

for

8w

kPr

ogre

ssio

nto

war

dth

ego

alof

10co

ntra

ctio

nsof

10s

dura

tion

(1–2

wk)

.Pr

ogre

ssio

nto

func

tiona

lact

iviti

esw

hen

pat

ient

sw

ere

able

to:

(1)

cont

ract

mus

cle

ina

spec

ific

pat

tern

and

(2)

per

form

10co

ntra

ctio

nsof

10s

hold

s(3

–5w

k).

Hea

vier

-load

func

tiona

ltas

ksw

ere

pro

gres

sive

lyin

trod

uced

inth

ela

st3

wk

ofth

ep

rogr

am.

Hom

eex

erci

ses

incl

uded

Adh

eren

cew

as12

.12

(2.6

9)se

ssio

nsp

erp

atie

nt,

and

hom

eex

erci

ses

had

med

ian

of23

.5se

ssio

ns

Tact

ilean

dp

ress

ure

cues

Mill

eret

al,6

120

056

wk

Trea

tmen

tw

asdi

vide

din

to3

pha

ses.

Phas

e1

goal

was

top

erfo

rm10

rep

etiti

ons

of10

-sho

lds

indi

ffere

ntp

ositi

ons.

Phas

e2

goal

was

cont

ract

ion

ofth

etr

ansv

ersu

sab

dom

inis

and

mul

tifidu

sm

uscl

esw

ithlo

adin

gof

the

limbs

indi

ffere

ntp

ositi

ons.

Phas

e3

goal

was

mor

eco

mp

lex

load

ing

exer

cise

s.

Patie

nts

wer

eas

ked

top

erfo

rmap

pro

xim

atel

y10

–15

min

ofho

me

exer

cise

s

Not

stat

edVe

rbal

,ta

ctile

,an

dp

ress

ure

gaug

e

Mos

eley

,27

2002

2se

ssio

nsp

erw

eek

for

4w

kN

otst

ated

Stan

dard

hom

eex

erci

ses

Not

stat

edN

otst

ated

Nie

mis

toet

al,6

2

2003

1se

ssio

np

erw

eek

for

4w

kPr

ogre

ssio

nw

asp

erfo

rmed

byin

stru

ctin

gth

ep

atie

nts

top

erfo

rmex

erci

ses

ina

mor

e-fu

nctio

nalm

anne

ran

dfu

rthe

rin

tegr

ate

them

inda

ilyac

tiviti

es.

Verb

al,

visu

al,

tact

ile,

and

pre

ssur

ega

uge

(Con

tinue

d)



Tab

le2.

Con

tinue

d

Art

icle

Du

rati

on

of

Mo

tor

Co

ntr

ol

Inte

rven

tio

nP

rog

ress

ion

Ru

leH

om

eP

rog

ram

Ad

her

ence

Mea

n(S

D)

Feed

bac

k

O’S

ulliv

anet

al,1

4

1997

1se

ssio

np

erw

eek

for

10w

kH

oldi

ngtim

eof

exer

cise

sw

asin

crea

sed

grad

ually

,as

wel

las

the

pre

ssur

eon

biof

eedb

ack

mon

itor.

Goa

lwas

10co

ntra

ctio

nsof

10-s

hold

s.Fu

rthe

rlo

wlo

ads

wer

eap

plie

dby

addi

ngle

vera

geth

roug

hlim

bs.

Whe

nac

cura

teac

tivat

ion

ofth

eco

-co

ntra

ctio

np

atte

rnw

asac

hiev

ed,

exer

cise

sw

ere

pro

gres

sed

tofu

nctio

nalh

oldi

ngof

pos

ture

san

dac

tiviti

eskn

own

top

revi

ousl

yag

grav

ate

pat

ient

s’sy

mp

tom

s.

Patie

nts

wer

eas

ked

todo

daily

exer

cise

sof

app

roxi

mat

ely

10–1

5m

in

Patie

nts

com

ple

ted

ada

ilyex

erci

ses

shee

tto

mon

itor

adhe

renc

e,bu

tre

sults

wer

eno

tp

rese

nted

Pres

sure

gaug

e

Rasm

usse

n-Ba

rret

al,6

620

031

sess

ion

of45

min

per

wee

kfo

r6

wk

Exer

cise

sw

ere

pro

gres

sed

byap

ply

ing

low

load

toth

em

uscl

eth

roug

hth

elim

bsin

diffe

rent

pos

ition

s.Pa

tient

sw

ere

inst

ruct

edin

how

tous

eco

ntra

ctio

nof

the

mus

cles

durin

gac

tiviti

esof

daily

livin

gan

din

situ

atio

nsth

atse

tof

fp

ain.

Patie

nts

wer

eas

ked

todo

daily

exer

cise

sof

app

roxi

mat

ely

10–1

5m

in

Not

stat

edTa

ctile

and

pre

ssur

ega

uge

Shau

ghne

ssy

etal

,63

2004

10se

ssio

nsin

10w

kTh

isco

nsis

ted

oftw

o1-

hse

ssio

nsdu

ring

wee

k1,

two

30-m

inse

ssio

nsdu

ring

wee

k2,

one

30-m

inse

ssio

ndu

ring

each

ofw

eeks

3–6,

and

one

30-m

inse

ssio

ndu

ring

wee

ks8

and

10.

Con

trac

tions

wer

efir

stp

erfo

rmed

with

the

goal

toac

hiev

e10

cont

ract

ions

of10

-sho

lds.

Onc

ep

atie

nts

wer

eab

leto

per

form

sust

aine

dco

ntra

ctio

nsin

low

-load

pos

ture

s,th

ere

gim

enw

asp

rogr

esse

dby

addi

ngle

vera

geth

roug

hlim

bm

ovem

ents

.

Patie

nts

per

form

edda

ilym

aint

enan

ceex

erci

ses

atho

me

Not

stat

edVe

rbal

,vi

sual

,ta

ctile

,an

dp

ress

ure

gaug

e

Stev

ens

etal

,24

2007

18in

divi

dual

sess

ions

of45

min

in12

wk

(2tim

esp

erw

eek

inth

efir

st6

wk

and

1tim

ep

erw

eek

inth

ene

xt6

wk)

Exer

cise

sw

ere

pra

ctic

edin

diffe

rent

envi

ronm

ents

and

cont

exts

tom

axim

ize

tran

sfer

sto

daily

situ

atio

ns.

The

phy

sica

lth

erap

ist

was

free

toch

oose

the

typ

eof

exer

cise

and

the

pro

gres

sion

hefe

ltm

ost

suita

ble

for

indi

vidu

alp

atie

nt.

Base

don

cont

inuo

uscl

inic

alex

amin

atio

n,th

etr

eatm

ent

pro

cess

cont

aine

da

clea

rlin

eof

pro

gres

sion

achi

eved

bych

angi

ngp

aram

eter

ssu

chas

pos

tura

lloa

d,re

duct

ion

ofat

tent

ion

dem

ands

,re

duct

ion

ofsp

eed,

orad

ditio

nals

trat

egie

sto

augm

ent

per

form

ance

,w

ithth

efin

algo

alto

obta

infu

nctio

nali

mp

rove

men

t.

Dai

lyho

me

exer

cise

sw

ere

enco

urag

ed;

how

ever

,ad

here

nce

was

not

asse

ssed

Not

stat

edN

otst

ated

Stug

eet

al,2

820

04Se

ssio

nsof

30to

60m

in,

3da

ysp

erw

eek,

for

18to

20w

k

Firs

t,th

efo

cus

was

onth

esp

ecifi

cco

ntra

ctio

nof

the

tran

sver

sely

orie

nted

abdo

min

alm

uscl

e.A

fter

app

roxi

mat

ely

4w

k,lo

adin

gw

asp

rogr

essi

vely

incr

ease

d.

Exer

cise

sw

ere

mai

nly

per

form

edat

hom

e.Pa

tient

sw

ere

enco

urag

edto

activ

ate

the

tran

sver

sus

abdo

min

ism

uscl

esre

gula

rlydu

ring

daily

activ

ities

.

Adh

eren

cew

as11

sess

ions

per

pat

ient

.80

%of

pat

ient

sdi

dth

eir

exer

cise

pro

gram

3tim

esp

erw

eek,

eith

erat

the

clin

icor

atho

me.

Not

stat

ed



Figure 2.Forest plot of the results of randomized controlled trials comparing motor control exercises with minimal intervention or motorcontrol exercises as a supplement. Values presented are effect size (weighted mean difference) and 95% confidence interval. Thepooled effect sizes were calculated using a random-effects model except for quality of life at intermediate and long-term follow-ups.



Figure 3.Forest plot of the results of randomized controlled trials comparing motor control exercises with spinal manipulative therapy. Valuesrepresent effect size (weighted mean difference) and 95% confidence interval. The pooled effect size was calculated using afixed-effect model.



effect was statistically significant.The results showed that motor con-trol exercise was better than otherforms of exercises only for reduc-ing disability at short-term follow-up(weighted mean difference��5.1points, 95% CI��8.7 to 1.4) (Fig. 4).The results of a single trial26 showedno difference between treatmentgroups for quality of life at short-term follow-up.

Motor Control Exercise VersusSurgeryOnly one study68 compared motorcontrol exercise with surgery, with amethodological quality score of 8.Surgery consisted of lumbar fusionwith transpedicular screws of theL4–L5 segments or the L5–S1 seg-ments. Brox et al68 found no statisti-cally significant differences for pain(mean difference [on a 0–100scale]��9 points, 95% CI��22.1 to3.5), disability (mean difference�

�3.3 points, 95% CI��12.8 to 6.2),and quality of life (mean difference�0.4 points, 95% CI��1.6 to 0.8) atthe long-term follow-up (Fig. 5).

DiscussionThis systematic review provides evi-dence that motor control exercise,alone or as a supplement to anothertherapy, is effective in reducing painand disability in patients with persis-tent, nonspecific LBP. We did notfind convincing evidence that motor

Figure 4.Forest plot of the results of randomized controlled trials comparing motor control exercises with other forms of exercise. Valuesrepresent effect size (weighted mean difference) and 95% confidence interval. The pooled effect size was calculated using arandom-effects model for pain at short-term follow-up and for disability at long-term follow-up and using a fixed-effect model forall other comparisons.



control exercise was superior tomanual therapy, other forms of exer-cise, or surgery.

Figure 2 shows that there was somevariation among studies in the effectsizes for motor control exercise. Fea-tures that could influence the treat-ment effect sizes are characteristicsof the patients (eg, symptom dura-tion), characteristics of treatment im-plementation (eg, program duration,experience of the therapist), and themethodological quality of the trial.Unfortunately, there are too few tri-als to systematically evaluate the ef-fects of these features using tech-niques such as meta-regression.

An intriguing finding of this reviewwas that motor control exercise wasas effective in reducing pain and in-creasing quality of life as a less-complex form of exercise therapythat did not incorporate the retrain-ing of specific muscles that often istime consuming to therapists and pa-tients. When taking in considerationthe results for disability, motor con-trol exercise was more effective thanother forms of exercise only at short-term follow-up, but the point esti-mate was small (5.1 out of 100),showing differences between inter-ventions that may not be clinicallyimportant.

The results of a single trial68 showedthat motor control exercise was notmore effective than surgery. Thisfinding is interesting because bothinterventions target the restoration

of spinal stability, and although spi-nal stability was not directly mea-sured, the findings suggest that themotor control approach is as effec-tive in maintaining stability as an in-vasive intervention that creates sta-bility by fusing the spine. However,this was the finding of a single trial,and more research is needed to con-firm the results.

Although a motor control interven-tion has been shown to reduce pain,it is still unknown whether thesechanges are accompanied by im-provements in measures of motorcontrol. Tsao and Hodges69 haveshown improvements in motor con-trol (anticipatory contraction of thetransversus abdominis muscle duringarm movement) after a single treat-ment session where the isolation ofthe transversus abdominis musclewas trained. In a different trial, Halland colleagues70 did not find thatmotor control (anticipatory contrac-tion of the transversus abdominismuscle during arm movement and awalking task) changed after trainingthe trunk muscles in a nonisolatedmanner. Therefore, the results ofthese 2 studies support the princi-ples of a motor control interventionwhere the isolated training of thedeep trunk muscles is emphasized.However, there has not been a pub-lished randomized controlled trialthat used clinical and physiologicalmeasures to detect improvements inmotor control that can be associatedwith improvements in pain and dis-

ability and the maintenance of thesechanges.

One question that is still to be an-swered is whether individuals withreduced motor control respond bestto this intervention or whether thereare other clinical features that can beused to define a subgroup of patientswho will respond best to this type ofintervention.

A standard protocol and definitionsfor motor control exercise are yet tobe established, and this is reflectedin the wide variation among trials inhow the exercise was named andimplemented (Tab. 2). Although inmost cases O’Sullivan et al14 and Ri-chardson et al71 were cited as refer-ences, it is apparent from inspectionof the articles that the interventionsin the trials were quite heteroge-neous. There was variation in theduration of the exercise program,progression rule, use of home exer-cise programs, and type of feedbackused with the motor control inter-vention. As an illustration, the pro-gram lasted 10 weeks in the trial byO’Sullivan et al, whereas the pro-gram lasted 18 to 20 weeks in thetrial by Stuge et al.28 In the trial byFerreira et al,13 ultrasound was usedfor feedback, and Stuge et al28 usedTerapi Master exercise equipment†:2 elements missing from the trial byO’Sullivan and colleagues.

† Nordisk Terapi A/S, Kilsund 4290, Staubo,Norway.

Figure 5.Forest plot of the results of a randomized controlled trials comparing motor control exercises with surgery. Values represent meandifference and 95% confidence interval.



Detailed comparison among trials isdifficult because in many trials theauthors did not thoroughly describethe motor control intervention thatwas evaluated. Accordingly, al-though we can conclude from thisreview that motor control exercise isan effective treatment for persistentLBP, the optimal way to implementthis intervention is not yet clear.

When looking at the quality of thetrials included in this review, a meanscore of 6 can be considered a highscore because these trials were exer-cise trials where it is impossible toblind the treatment provider andsubjects, and, therefore, the maxi-mum PEDro score that can beachieved is 8. However, becausesome trials were of lower method-ological quality, they potentiallypresent biased (and overly optimis-tic) estimates of treatment effects.To assess the impact of the lower-quality studies on the review conclu-sions, a sensitivity analysis with ex-clusion of trials with scores lowerthan 524,64 was performed. When thelower-quality studies were deleted,the effect size unexpectedly in-creased slightly for pain and disabil-ity outcomes (we did not conduct asensitivity analysis for quality of lifebecause the exclusion of these trialswould leave only one trial in thetreatment contrast). Therefore, wedo not believe that our conclusionthat motor control exercise is effec-tive (compared with minimal inter-vention or when used as a supple-ment) is an artifact of the inclusionof low-quality trials.

This review not only includes 4 newtrials that were not included in pre-vious reviews, accounting for theaddition of 560 patients, but also al-lowed the use of a meta-analyticalapproach with the inclusion of agreater number of articles into eachtreatment contrast. The pooled re-sults of this systematic reviewshowed smaller and more-precise es-

timates of treatment effects whencompared with the pooled results ofFerreira et al.13 This differenceamong studies can be seen whenlooking, for example, at the motorcontrol exercise versus minimal in-tervention contrast. For this con-trast, Ferreira et al13 included 2 trialsand found an effect of �21 on a 0 to100 scale (95% CI��32 to �9) forpain, whereas we found, based on 5trials, an effect of �14.3 (95%CI��20.4 to �8.1).

Although it has been only recentlythat reviews of motor control exer-cises have been published, this typeof intervention is widely acceptedand used in the clinical field aroundthe world. Therefore, it is still crucialthat further studies in the area bedeveloped, such as a placebo-controlled trial and trials aiming toidentify subgroups of patients whowill benefit more from a motor con-trol intervention. More fundamentalstudies in LBP to establish reliableand valid clinical assessment tools toidentify deficits in motor control alsoare needed.

ConclusionThe results of this systematic reviewsuggest that motor control exerciseis more effective than minimal inter-vention and adds benefit to anotherform of intervention in reducingpain and disability for people withpersistent LBP. The optimal imple-mentation of motor control exerciseat present is unclear. Future trialsevaluating issues such as dosage pa-rameters, feedback approaches, andeffects in defined subgroups are ahigh priority.

Ms Macedo, Dr Maher, and Dr Latimer pro-vided concept/idea/research design anddata collection. Ms Macedo and Dr Maherprovided writing and data analysis. MsMacedo, Dr Maher, and Dr McAuley pro-vided project management. Dr Latimer pro-vided clerical support and consultation (in-cluding review of manuscript beforesubmission).

Ms Macedo holds a PhD scholarship jointlyfunded by The University of Sydney and theAustralian Government. Dr Maher’s researchfellowship is funded by Australia’s NationalHealth and Medical Research Council.

This article was received April 3, 2008, andwas accepted October 10, 2008.

DOI: 10.2522/ptj.20080103

References1 Hancock MJ, Maher CG, Latimer J, et al.

Systematic review of tests to identify thedisc, SIJ or facet joint as the source oflow back pain. Eur Spine J. 2007;16:1539–1550.

2 Niemisto L, Rissanen P, Sarna S, et al. Cost-effectiveness of combined manipulation,stabilizing exercises, and physician consul-tation compared to physician consultationalone for chronic low back pain: a pro-spective randomized trial with 2-yearfollow-up [with consumer summary].Spine. 2005;30:1109–1115.

3 Niemisto L, Sarna S, Lahtinen-Suopanki T,et al. Predictive factors for 1-year outcomeof chronic low back pain following manip-ulation, stabilizing exercises, and physi-cian consultation or physician consulta-tion alone. J Rehabil Med. 2004;36:104–109.

4 Panjabi MM. Clinical spinal instability andlow back pain. J Electromyogr Kinesiol.2003;13:371–379.

5 Hodges PW, Richardson CA. Delayed pos-tural contraction of transversus abdominisin low back pain associated with move-ment of the lower limb. J Spinal Disord.1998;11:46–56.

6 Hodges PW, Richardson CA. Relationshipbetween limb movement speed and asso-ciated contraction of the trunk muscles.Ergonomics. 1997;40:1220–1230.

7 MacDonald D, Moseley GL, Hodges PW.The function of the lumbar multifidus inunilateral low back pain. Presented at:World Congress of Low Back and PelvicPain; 2004; Melbourne, Australia.

8 Moseley GL, Hodges PW, Gandevia S.Deep and superficial fibers of the lumbarmultifidus muscle are differentially activeduring voluntary arm movements. Spine.2002;27:E29–E36.

9 Hodges PW, Richardson CA. Inefficientmuscular stabilisation of the lumbar spineassociated with low back pain: a motorcontrol evaluation of transversus abdomi-nis. Spine. 1996;21:2640–2650.

10 Hides JA, Stokes MJ, Saide M, et al. Evi-dence of lumbar multifidus muscle wast-ing ipsilateral to symptoms in patientswith acute/subacute low back pain. Spine.1994;19:165–177.

11 Roy SH, DeLuca CJ, Casavant DA. Lumbarmuscle fatigue and chronic low back pain.Spine. 1989;14:992–1001.

12 Alaranta H, Tallroth K, Soukka A, et al. Fatcontent of lumbar extensor muscles inlow back disability: a radiographic andclinical comparison. J Spinal Disord.1993;6:137–140.



13 Ferreira ML, Ferreira PH, Latimer J, et al.Comparison of general exercise, motorcontrol exercise and spinal manipulativetherapy for chronic low back pain: a ran-domized trial. Pain. 2007;131:31–37.

14 O’Sullivan PB, Phyty GD, Twomey LT,et al. Evaluation of specific stabilizing ex-ercise in the treatment of chronic lowback pain with radiologic diagnosis ofspondylolysis or spondylolisthesis. Spine.1997;22:2959–2967.

15 Hides JA, Richardson CA, Jull GA. Multifi-dus muscle recovery is not automatic afterresolution of acute, first-episode low backpain. Spine. 1996;21:2763–2769.

16 Ferreira PH, Ferreira ML, Maher CG, et al.Specific stabilisation exercise for spinaland pelvic pain: a systematic review. AustJ Physiother. 2006;52:79–88.

17 Hauggaard A, Persson A. Specific spinalstabilisation exercises in patients with lowback pain: a systematic review. Phys TherRev. 2007;12:233–248.

18 Rackwitz B, de Bie R, Limm H, et al. Seg-mental stabilizing exercises and low backpain. What is the evidence? A systematicreview of randomized controlled trials.Clin Rehabil. 2006;20:553–567.

19 Bombardier C, van Tulder MW, Pennick V,et al. Cochrane Back Group. About theCochrane Collaboration (Cochrane Re-view Groups [CRGs]). 2006:4.

20 Airaksinen O, Brox JI, Cedraschi C, et al.Chapter 4: European guidelines for themanagement of chronic nonspecific lowback pain. Eur Spine J. 2006;15(Suppl2):S192–S300.

21 van Tulder MW, Becker A, Bekkering T,et al. Chapter 3: European guidelines forthe management of acute nonspecific lowback pain in primary care. Eur Spine J.2006;15(Suppl 2):S169–S191.

22 Maher CG, Sherrington C, Herbert RD,et al. Reliability of the PEDro scale for rat-ing quality of randomized controlled trials.Phys Ther. 2003;83:713–721.

23 Sherrington C, Herbert RD, Maher C, et al.PEDro: a database of randomised con-trolled trials and systematic reviews inphysiotherapy. Man Ther. 2000;5:223–226.

24 Stevens V, Crombez G, Parlevliet T, et al.The effectiveness of specific exercise ther-apy versus device exercise therapy in thetreatment of chronic low back pain pa-tients. In: Proceedings of the 6th Interdis-ciplinary World Congress of Low Back andPelvic Pain; 2007; Barcelona, Spain; 2007:177.

25 Higgins J, Green S. Cochrane Handbookfor Systematic Reviews of Interventions4.2.6 [updated September 2006]. In: TheCochrane Library, issue 4, 2006. Chiches-ter, United Kingdom: John Wiley & SonsLtd; 2006.

26 Critchley DJ, Ratcliffe J, Noonan S, et al.Effectiveness and cost-effectiveness ofthree types of physiotherapy used to re-duce chronic low back pain disability: apragmatic randomized trial with eco-nomic evaluation [with consumer summa-ry]. Spine. 2007;32:1474–1481.

27 Moseley L. Combined physiotherapy andeducation is efficacious for chronic lowback pain. Aust J Physiother. 2002;48:297–302.

28 Stuge B, Laerum E, Kirkesola G, et al. Theefficacy of a treatment program focusingon specific stabilizing exercises for pelvicgirdle pain after pregnancy: a randomizedcontrolled trial. Spine. 2004;29:351–359.

29 van Tulder MW, Furlan A, Bombardier C,et al. Updated method guidelines for sys-tematic reviews in Cochrane Collabora-tion Back Review Group. Spine. 2003;28:1290–1299.

30 Bendix AF, Bendix T, Lund C, et al. Com-parison of three intensive programs forchronic low back pain patients: a prospec-tive, randomized, observer-blinded studywith one-year follow-up. Scand J RehabilMed. 1997;29:81–89.

31 Bentsen H, Lindgarde F, Manthorpe R. Theeffect of dynamic strength back exerciseand/or a home training program in 57-year-old women with chronic low backpain: results of a prospective randomizedstudy with a 3-year follow-up period.Spine. 1997;22:1494–1500.

32 Cambron JA, Gudavalli MR, Hedeker D,et al. One-year follow-up of a randomizedclinical trial comparing flexion distractionwith an exercise program for chronic low-back pain. J Altern Complement Med.2006;12:659–668.

33 Friedrich M, Gittler G, Halberstadt Y, et al.Combined exercise and motivation pro-gram: effect on the compliance and levelof disability of patients with chroniclow back pain—a randomized controlledtrial. Arch Phys Med Rehabil. 1998;79:475–487.

34 Frost H, Lamb SE, Doll HA, et al. Random-ised controlled trial of physiotherapy com-pared with advice for low back pain. BMJ.2004;329:708–713.

35 Gudavalli MR, Cambron JA, McGregor M,et al. A randomized clinical trial and sub-group analysis to compare flexion-distraction with active exercise forchronic low back pain. Eur Spine J.2006;15:1070–82.

36 Helewa A, Goldsmith CH, Lee P, et al.Does strengthening the abdominal mus-cles prevent low back pain: a randomizedcontrolled trial. J Rheumatol. 1999;26:1808–1815.

37 Koes BW, Bouter LM, van Mameren H,et al. The effectiveness of manual therapy,physiotherapy, and treatment by the gen-eral practitioner for nonspecific back andneck complaints: a randomized clinical tri-al. Spine. 1992;17:28–35.

38 Lie H, Frey S, Lie H, et al. Mobilizing orstabilizing exercise in degenerative diskdisease in the lumbar region? [in Norwe-gian]. Tidsskrift for Den Norske Laege-forening. 1999;119:2051–2053.

39 Mannion AF, Muntener M, Taimela S, et al.Comparison of three active therapies forchronic low back pain: results of a ran-domized clinical trial with one-year follow-up. Rheumatology. 2001;40:772–778.

40 Nelson BW, O’Reilly E, Miller M, et al. Theclinical effects of intensive, specific exer-cise on chronic low back pain: a con-trolled study of 895 consecutive patientswith 1-year follow up. Orthopedics. 1995;18:971–981.

41 Shaughnessy A. Can a specific exerciseprogram combined with brief counselingby a physical therapist offer benefits overusual care? Evidence-Based Practice.1999;2:1–2.

42 Suni J, Rinne M, Natri A, et al. Control ofthe lumbar neutral zone decreases lowback pain and improves self-evaluatedwork ability: a 12-month randomized con-trolled study. Spine. 2006;31:E611–E620.

43 Timm KE. A randomized-control study ofactive and passive treatments for chroniclow back pain following L5 laminectomy.J Orthop Sports Phys Ther. 1994;20:276–286.

44 Freburger JK, Carey TS, Holmes GM, et al.Effectiveness of physical therapy for themanagement of chronic spine disorders: apropensity score approach. Phys Ther.2006;86:381–394.

45 Hicks GE, Fritz JM, Delitto A, et al. Prelim-inary development of a clinical predictionrule for determining which patients withlow back pain will respond to a stabiliza-tion exercise program. Arch Phys Med Re-habil. 2005;86:1753–1762.

46 Kasai R. Current trends in exercise man-agement for chronic low back pain: com-parison between strengthening exerciseand spinal segmental stabilization exer-cise. J Phys Med Sci. 2006;18:97–105.

47 Ljungkvist I, Ljungkvist I. Short- and long-term effects of a 12-week intensive func-tional restoration programme in individu-als work-disabled by chronic spinal pain.Scand J Rehabil Med Suppl. 2000;40:1–14.

48 Weinhardt C, Heller KD, Weh L, et al.Non-operative treatment of chronic lowback pain: specific back muscular strengthtraining versus improvement of physicalfitness. Zeitschrift fur Orthopadie undIhre Grenzgebiete. 2001;139:490–495.

49 Koumantakis GA, Watson PJ, Oldham JA,et al. Supplementation of general endur-ance exercise with stabilisation trainingversus general exercise only: physiologicaland functional outcomes of a randomisedcontrolled trial of patients with recurrentlow back pain. Clin Biomech. 2005;20:474–482.

50 Stuge B, Veierod MB, Laerum E, et al. Theefficacy of a treatment program focusingon specific stabilizing exercises for pelvicgirdle pain after pregnancy: a two-yearfollow-up of a randomized clinical trial.Spine. 2004;29:E197–E203.

51 Cairns MC, Foster NE, Wright C, et al. Ran-domized controlled trial of specific spinalstabilization exercises and conventionalphysiotherapy for recurrent low backpain. Spine. 2006;31:E670–E681.

52 Gagnon LH. Efficacy of Pilates Exercisesas Therapeutic Intervention in TreatingPatients With Low Back Pain [disserta-tion]. Knoxville, TN: University of Tennes-see; 2005:119.



53 Hides JA, Jull GA, Richardson CA. Long-term effects of specific stabilizing exer-cises for first-episode low back pain.Spine. 2001;26:E243–E248.

54 Aure OF, Nilse JH, Vasseljen O. Manualtherapy and exercise therapy in patientswith chronic low back pain. Spine.2003;28:525–532.

55 Monticone M, Barbarino A, Testi C, et al.Symptomatic efficacy of stabilizing treat-ment versus laser therapy for sub-acutelow back pain with positive tests for sac-roiliac dysfunction: a randomized clinicalcontrolled trial with 1-year follow-up. Eu-ropa Medicophysica. 2004;40:263–268.

56 Lewis JS, Hewitt JS, Billington L, et al. Arandomized clinical trial comparing twophysiotherapy interventions for chroniclow back pain. Spine. 2005;30:711–721.

57 Riipinen M, Niemisto L, Lindgren KA, et al.Psychosocial differences as predictors forrecovery from chronic low back pain fol-lowing manipulation, stabilizing exercisesand physician consultation or physicianconsultation alone. J Rehabil Med. 2005;37:152–158.

58 Jull GA, Trott P, Potter H, et al. A random-ized controlled trial of exercises and ma-nipulative therapy for cervicogenic head-ache. Spine. 2002;27:1835–1843; discus-sion 1843.

59 Danneels LA, Cools AM, VanderstraetenGG, et al. The effects of three differenttraining modalities on the cross-sectionalarea of the paravertebral muscles. ScandJ Med Sci Sports. 2001;11:335–341.

60 Danneels LA, Vanderstraeten GG, CambierDC, et al. Effects of three different trainingmodalities on the cross-sectional area ofthe lumbar multifidus muscle in patientswith chronic low back pain. Br J SportsMed. 2001;35:186–191; comment in 2001;35:186–191.

61 Miller ER, Schenk RJ, Karnes JL, et al. Acomparison of the McKenzie approach toa specific spine stabilization program forchronic low back pain. Journal of Man-ual and Manipulative Therapy. 2005;13:103–112.

62 Niemisto L, Lahtinen-Suopanki T, RissanenP, et al. A randomized trial of combinedmanipulation, stabilizing exercises, andphysician consultation compared to phy-sician consultation alone for chronic lowback pain. Spine. 2003;28:2185–2191.

63 Shaughnessy M, Caulfield B, ShaughnessyM, et al. A pilot study to investigate theeffect of lumbar stabilisation exercisetraining on functional ability and quality oflife in patients with chronic low backpain. Int J Rehabil Res. 2004;27:297–301.

64 Goldby LJ, Moore AP, Doust J, et al. A ran-domized controlled trial investigating theefficiency of musculoskeletal physiother-apy on chronic low back disorder. Spine.2006;31:1083–1093.

65 Koumantakis GA, Watson PJ, Oldham JA.Trunk muscle stabilization training plusgeneral exercise versus general exerciseonly: randomized controlled trial of pa-tients with recurrent low back pain. PhysTher. 2005;85:209–225.

66 Rasmussen-Barr E, Nilsson-Wikmar L,Arvidsson I, et al. Stabilizing training com-pared with manual treatment in sub-acuteand chronic low-back pain. Man Ther.2003;8:233–241.

67 Kladny B, Fischer FC, Haase I, et al. Eval-uation of specific stabilizing exercise inthe treatment of low back pain and lumbardisk disease in outpatient rehabilitation.Zeitschrift fur Orthopadie und Ihre Gren-zgebiete. 2003;141:401–405.

68 Brox JI, Sørensen R, Friis A, et al. Random-ized clinical trial of lumbar instrumentedfusion and cognitive intervention and ex-ercise in patients with chronic low backpain and disc degeneration. Spine. 2003;28:1913–1921; comment in 2004;29:1913–1921.

69 Tsao H, Hodges PW. Immediate changes infeedforward postural adjustment follow-ing voluntary motor training. Exp BrainRes. 2007;181:537–546.

70 Hall L, Tsao T, MacDonald D, et al. Imme-diate effects of co-contraction training onmotor control of the trunk muscles in peo-ple with recurrent low back pain. J Elec-tromyogr Kinesiol. 2007 Nov 21 [Epubahead of print].

71 Richardson CA, Jull GA, Hodges PW. Ther-apeutic Exercise for Spinal SegmentalStabilization in Low Back Pain. Edin-burgh, Scotland: Churchill Livingstone;1999.



Documents

Motor Control Exercise for Persistent, Nonspecific Low Back Pain: A Systematic Review