Do Changes in Transversus Abdominis and Lumbar Multifidus During Conservative Treatment Explain Changes in Clinical Outcomes Related to Nonspecific Low Back Pain? A Systematic Review

The Journal of Pain, Vol -, No - (-), 2014: pp 1.e1-35Available online at www.jpain.org and www.sciencedirect.com

Online Exclusive

Do Changes in Transversus Abdominis and Lumbar Multifidus

During Conservative Treatment Explain Changes in Clinical

Outcomes Related to Nonspecific Low Back Pain?

A Systematic Review

Arnold Y. L. Wong,* Eric C. Parent,*,y Martha Funabashi,* and Gregory N. Kawchuk**Department of Physical Therapy, University of Alberta, Edmonton, Alberta, Canada.yGlenrose Rehabilitation Hospital, Edmonton, Alberta, Canada.

Arnold WGraduateGreg KawThere areAddressPhysical Tberta T6G

1526-590

ª 2014 b

http://dx

Abstract: Previous research describes an inconsistent relation between temporal changes in trans-

versus abdominis or lumbar multifidus and temporal changes in clinical outcomes. Unfortunately, a

relevant systematic review is unavailable. As a result, this systematic review was designed to summa-

rize evidence regarding the association between temporal changes in muscle morphometry and ac-

tivity in response to treatment and temporal changes in clinical outcomes. Candidate publications

were identified from 6 electronic databases. Fifteen articles were included after scrutinization by 2

reviewers using predetermined selection criteria. The methodological quality of these articles was

appraised using a standard tool. These methods revealed strong evidence that temporal alterations

in transversus abdominis thickness change during contraction (as measured by B-mode or M-mode

ultrasound) or feedforward activation of transversus abdominis (assessed via electromyography, tis-

sue Doppler imaging, or M-mode ultrasound) were unrelated to temporal changes in low back pain

(LBP)/LBP-related disability. There was limited evidence that temporal changes in transversus abdom-

inis lateral sliding or lumbar multifidus endurance were unrelated to temporal changes in LBP inten-

sity. Conflicting evidence was found for the relation between temporal changes in lumbar multifidus

morphometry and temporal changes in LBP/LBP-related disability. This review highlights that tempo-

ral changes in transversus abdominis features tend to be unrelated to the corresponding LBP/LBP-

related disability improvements, whereas the relation between multifidus changes and clinical

improvements remains uncertain.

Perspective: This systematic review highlighted that changes in morphometry or activation of

transversus abdominis following conservative treatments tend not to be associated with the corre-

sponding changes in clinical outcomes. The relation between posttreatment changes in characteris-

tics of lumbar multifidus and clinical improvements remains uncertain.

ª 2014 by the American Pain Society

Key words: Transversus abdominis, lumbar multifidus, temporal changes, low back pain, ultrasound

imaging.

Long is supported by the Alberta Innovates-Health SolutionsStudentship and the Golden Key Graduate Scholar Award.chuk is supported by the Canadian Research Chair Program.no conflicts of interest.

reprint requests to Greg N. Kawchuk, DC, PhD, Department ofherapy, University of Alberta, 2-40 Corbett Hall, Edmonton, Al-2G4, Canada. E-mail: [email protected]

0/$36.00

y the American Pain Society

.doi.org/10.1016/j.jpain.2013.10.008

ow back pain (LBP) is a commonmusculoskeletal dis-order55 with a reported lifetime prevalence ofapproximately 80%.52,106 Although 70 to 90% of

LBP patients will recover in 2 to 6 weeks,15,105

approximately 60 to 86% of patients with a firstepisode of LBP will relapse within a year,64 whereas 6to 10% of patients develop chronic LBP.82 The high prev-alence of LBP is associatedwith enormous socioeconomicburdens22,71 and disability.104 Notwithstanding currentdiagnostic technology, the cause of LBP remains

1.e1

Figure 1. A flow diagram of the literature search.

1.e2 The Journal of Pain Deep Trunk Muscles and Back Pain—A Systematic Review

unknown in approximately 90% of patients (ie, nonspe-cific LBP),34 which in turn hinders effective treatmentand prevention of LBP.Although the precise cause of LBP remains elusive,

previous research has found that patients with LBPdemonstrated neuromuscular, morphometric, or histo-logic changes in transversus abdominis (TrA) orlumbar multifidus (LM).6,18,24,26,43,47,48,51,56,59,69,72,93,107

Compared with asymptomatic individuals, patientswith LBP displayed reduced TrA/LM thickness changeduring contraction,16,19,58,59,107 delayed feedforwardactivation of TrA/deep LM during trunk or limbmovement,45,46,70,75,84 and fat infiltration and muscleatrophy in LM.81,84 Given the anatomic positions of TrA/LM8,28,40,54 and the impairments of these muscles in LBPpatients, TrA/LM are postulated to play an importantrole in maintaining intervertebral stiffness/stability andpreventing LBP recurrence.28,41

Specific conservative interventions have been de-signed to restore the morphometry, histology, and acti-vation of TrA/LM in patients with LBP, which mayimprove the clinical outcomes and prevent LBP recur-rence.77,85,97 In other words, these variables aretargeted by various treatments and are thereforehypothesized to create changes in clinical outcomes.108

Although a prior systematic review has summarizedthat baseline characteristics of TrA/LM may not predictfuture clinical outcomes of LBP,115 there are conflictingfindings regarding the relation between temporalchanges in morphometry/feedforward activation ofthese variables and changes in clinical outcomes duringtreatment or at discharge. One study found thatincreased TrA and decreased obliquus internus contrac-tion thickness ratios measured with B-mode ultrasoundimaging (USI) in response to 8 weeks of various exerciseinterventions explained 18% of the variance in temporalLBP reduction.102 On the contrary, another study showedthat the improvement of LBP-related disability following

9-week motor control exercises was associated withneither temporal changes in TrA contraction ratio nortemporal changes in feedforward activation of deepabdominal muscles (including TrA) measured withUSI.73 Similarly, improvements in functional scores of pa-tients with LBP following 2 sessions of spinal manipula-tive therapy were unrelated to the correspondingalterations in LM percent thickness change measuredwith B-mode USI.31,79 Collectively, knowing whichvariables can predict treatment outcomes would helpclinicians apply TrA/LM-targeted treatments withimproved confidence and may help design interventionsto increase success. However, to our knowledge, a rele-vant systematic review has yet to be conducted.Given the above, the primary objective of this system-

atic review was to summarize published evidenceregarding an association between temporal changes inTrA/LM muscles treated with conservative interventionsand temporal changes in clinical outcomes of patientswith nonspecific LBP. The secondary objective was to re-view whether the relation between changes in TrA/LMand clinical outcomes was affected by age, gender, andLBP chronicity.

MethodsThe review protocol was registered with PROSPERO

(CRD42013003860). The review followed the suggestedmethodological and reporting guidelines of thePreferred Reporting Items of Systematic Reviews andMeta-analyses (PRISMA)78 and the Meta-analysis ofObservational Studies in Epidemiology (MOOSE).90

Identification of StudiesRelevant literature in English, Chinese, French, and

Portuguese was identified by a systematic search ofMEDLINE, EMBASE, PEDro, SPORTDiscus, CINAHL, andthe Cochrane Library (from inception to March 2013).

Table 1. Description of the Included Studies Classified Based on Dynamic and Static Morphometry of Transversus Abdominis and LumbarMultifidus

STUDYSTUDY POPULATION; AVERAGE AGE 6

STANDARD DEVIATION; FOLLOW-UP INTERVENTION

TEMPORAL CHANGE IN

MUSCLE VARIABLES

TEMPORAL

CHANGE IN

OUTCOME

MEASURE

COFOUNDER

ADJUSTMENT RESULTS/EXPLAINED VARIANCE

Studies investigating TrA dynamic morphometry or activation

Brook

et al, 201210,*

Participants with chronic LBP

(n = 64); Group A (n = 32):

36.2 6 8.2 years; Group B

(n = 32): 36.3 6 6.3 years;

8 weeks

A randomized controlled

trial. Each group received

8-week of 3 times per

week exercise.

A: Specific exercise group:

Pilates, ADIM, abdominal

bracing, full-body exercise,

whole-body stretching.

B: General exercise group:

cycling exercises with different

progression, sprint focus,

whole-body stretching

Feedforward activation

of TrA/OI as measured

by surface EMG during

rapid right shoulder

flexion

1. ODI

2. VAS

(current)

3. VAS (last

week)

No adjustment for

baseline TrA/OI

Results obtained from personal

communication

using pooled data:

No statistically significant

correlation was found

between changes in

anticipatory onset

of TrA/OI activation (either side)

and changes in:

1. r = �.05 and .02

2. r = �.08 and �.04

3. r = .00 and .06 (P value was

not reported).

Ferreira

et al, 201025Patients with chronic nonspecific

LBP recruited from physical

therapy departments at 3

teaching hospitals (n = 34);

Group A (n = 11):

45.4 6 17.3 years; Group B

(n = 10): 54.9 6 11.3 years;

Group C (n = 13):

45.4 6 17.7 years; 8 weeks

A randomized controlled trial.

Each group received 12

sessions over 8 weeks:

A: Motor control exercises:

training function of specific

deep muscles of lumbar region,

coordination of deep trunk

muscles and pelvic floor

muscles with diaphragmatic

respiration pattern,

incorporation of deep trunk

muscles contraction during

functional tasks, home

exercises

B: General exercises:

strengthening and stretching

of main muscle groups,

cardiovascular fitness,

education, home exercises

C: Spinal manipulative therapy:

joint mobilization but not

thrust manipulation techniques

to the spine or pelvis

TrA percent thickness

change during

an isometric knee

flexion/extension

task as measured by

B-mode USI

1. Global

impression

of recovery

2. RMDQ

3. PSFS score

4. VAS

Corresponding baseline

clinical outcomes

Results obtained from the pooled

sample. The correlation between

improvements in TrA percent

thickness change during

contraction and improvements

in the clinical outcomes

1. r = .27; 95% CI = �.08, .55

2. r = �.35; 95% CI = �.62, �.02

(increased TrA contraction has

lower RMDQ score)

3. r = .19; 95% CI = �.16, .50

4. r = �.28; 95% CI = �.56, .07

Wonget

alTheJournalofPain

1.e3

Table 1. Continued

STUDYSTUDY POPULATION; AVERAGE AGE 6

STANDARD DEVIATION; FOLLOW-UP INTERVENTION

TEMPORAL CHANGE IN

MUSCLE VARIABLES

TEMPORAL

CHANGE IN

OUTCOME

MEASURE

COFOUNDER

ADJUSTMENT RESULTS/EXPLAINED VARIANCE

Konitzer

et al, 201160Participants with LBP who met at

least 3 out of 4 criteria for

lumbar stabilization CPR were

recruited from a physical

therapy department in an army

center (n = 19); 32.5 6 7.8 years;

3 to 4 days later

An uncontrolled study. Participants

received a single session of

standardized high-velocity, low-

amplitude spinal manipulative

therapy to their back, and

returned 3–4 days later for

NPRS, ODI, and USI assessment

TrA percent thickness change

during ADIM as measured

by B-mode USI

NPRS

ODI

No confounders

were adjusted for.

No significant correlations between

immediate post–spinal

manipulative therapy changes

in TrA percent thickness change

and changes in NPRS at the

first session; No significant

correlations between temporal

alterations in TrA percent

thickness change and temporal

changes in ODI scores at the

second session (P = .23 to P = .94)

Marshall

et al, 200677Patients with chronic LBP were

recruited from general medical

practitioners, gym goers,

chiropractors, physiotherapists,

and through television and

newspaper advertisements

(n = 20); 38.8 6 12.1 years; 4,

8, and 12 weeks

An uncontrolled study. Standard

progressive Swiss ball exercises

were performed 3 times per

week for 12 weeks

Feedforward activation

of bilateral TrA/OI as

measured by surface

EMG

ODI No adjustment for

confounders

No significant change in latency of

TrA/OI over the time course of

the experiment (P-value was

not reported).

Changes in feedforward activation

of TrA/OI were not retained in

the 2 multiple linear regression

models that explained the change

in ODI score at 8 weeks and

12 weeks, respectively

Mannion

et al, 201273Patients with chronic nonspecific

LBP and with or without

referred pain were recruited

from rheumatology,

orthopedics, and neurology of

participating hospitals (n = 37);

44 6 12.3 years; 9 weeks

An uncontrolled study. Each

patient received 1 motor control

exercise session per week for

9 weeks.

Motor control exercise: training

function of specific deep muscles

of lumbar region, coordination of

deep trunk muscles and pelvic

floor muscles with diaphragmatic

respiration pattern, incorporation

of deep trunk muscles contraction

during functional tasks,

education, home exercises

(10 � 10s repetition, 10 times

per day)

(i) Voluntary TrA contraction

thickness ratio as

measured by M-mode

USI during ADIM

(ii) Feedforward onset of

lateral abdominal muscles

(TrA, OI, OE) during rapid

arm movement measured

with TDI

RMDQ

Pain graphic

rating scale

No confounder

was adjusted for.

No significant correlation between

posttreatment change in (i) and

change in RMDQ scores (r = .08,

P = .66) or pain (actual r values and

P-values were not reported).

No significant correlation between

change in (ii) and change in RMDQ

scores (r = .16, P = .40) or pain (r

value and P-value were not

reported).

1.e4

TheJournalofPain

Deep

TrunkMuscles

andBack

Pain—

ASystem

aticReview

Table 1. Continued

STUDYSTUDY POPULATION; AVERAGE AGE 6

STANDARD DEVIATION; FOLLOW-UP INTERVENTION

TEMPORAL CHANGE IN

MUSCLE VARIABLES

TEMPORAL

CHANGE IN

OUTCOME

MEASURE

COFOUNDER

ADJUSTMENT RESULTS/EXPLAINED VARIANCE

Tsao and

Hodges,

200898

Volunteers with unilateral or

bilateral chronic LBP were

recruited from unknown

population (n = 9);

26 6 7 years; 1 week, 2 week,

4 week, 6 months

An uncontrolled trial. All participants

attended 2 sessions of motor

control exercise training in the first

and the second week.

The exercise included 3 sets of

10s � 10 repetitions of isolated

TrA contraction at 5%

maximum voluntary contraction.

Participants were instructed to

continue the exercise twice per

day for 4 weeks at home

Feedforward onset of TrA

during left rapid arm

flexion/extension

as measured by fine-wire

EMG and the onset timing

was detected by visual

inspection

1. VAS

2. PSFS

No confounder was

adjusted for (eg,

exercise compliance).

The correlation between temporal

changes in TrA onset timing at

4 weeks and the corresponding

changes in VAS and PSFS scores

1. Flexion r = �.10, P = .80;

Extension r = .10, P = .80

2. Flexion r = �.1, P = .68;

Extension r = .18, P = .64

Tsao

et al, 201096Right-handed volunteers with

nonspecific LBP for at least

3 months, or episodic

nonspecific LBP (at least 2

episodes of LBP within

6 months, with each episode

lasting at least 3 days) (n = 20);

Group A (n = 10): 24 6 8 years;

Group B (n = 10): 23 6 3 years;

2 weeks

A randomized controlled trial.

Participants learned one of the

following exercise in a single

session and then continued

the exercise at home for 2 weeks:

A: Motor control exercise: isolated

contraction of TrA, pelvic floor

muscles with real-time ultrasound

biofeedback and encouraged

to practice 3 sets of 10s � 10

repetitions twice per day

B: Self-paced walking exercise of

10 minutes twice per day

Feedforward onset of TrA

bilaterally during left rapid

arm flexion/extension as

measured by fine-wire EMG

and the onset timing was

detected by visual

inspection offline

1. VAS

2. PSFS

No confounder was

adjusted for (eg,

exercise compliance).

The correlation between temporal

changes in TrA onset timing at 2

weeks and the corresponding

changes in VAS and PSFS scores

1. r < .22, P > .36

2. r < .27, P > .41

Unsgaard-

Tondel et al,

2012100

(same cohort

as Vasseljen

and Fladmark,

2010,102 and

Vasseljen

et al, 2012103)

Participants with chronic

nonspecific LBP recruited from

general practitioners or

physiotherapists and by

advertising among staff at a

local hospital (n = 109); Group

A (n = 36): 40.9 6 11.5 years;

Group B (n = 36):

43.4 6 10.2 years; Group C

(n = 37): 36 6 10.3 years;

1 year (a follow-up

questionnaire on NPRS only)

A randomized controlled trial.

Each group received weekly

sessions over 8 weeks.

A: Motor control exercises

(40 minutes each): training

function of TrA, LM, and pelvic

floor muscle, cocontraction of

deep trunk muscles and pelvic

floor muscles, incorporation of

deep trunk muscles contraction

during functional tasks, home

exercises

B: General exercises (1 hour):

general trunk strengthening,

Abdominal muscle

function at the end of

8-week treatment:

(i) TrA lateral slide

during ADIM

(ii) TrA-CR during ADIM

(iii) OI contraction thickness

ratio during ADIM

(iv) Anticipatory onset of

abdominal muscles during

contralateral rapid arm

flexion

(i), (ii), and (iii) were

measured by B-mode USI

NPRS Age, initial FABQ-P,

gender, BMI, initial

pain intensity and

pain duration were

adjusted for.

Age, initial pain level

and pain duration

were adjusted for.

Results obtained from the pooled

sample of all treatment groups.

Linear regression of pain at 1-year

follow-up in relation to immediate

posttreatment changes in:

(i) b = �.02; 95% CI = �.13, .08;

R2 = .2%

(ii) b = �.01; 95% CI = �.09, .08;

R2 = .0%

(iv) b = .02 95% CI, .00, .05;

R2 = 3.8%

Linear regression of pain at 1-year

follow-up in relation to both

baseline

Wonget

alTheJournalofPain

1.e5

Table 1. Continued

STUDYSTUDY POPULATION; AVERAGE AGE 6

STANDARD DEVIATION; FOLLOW-UP INTERVENTION

TEMPORAL CHANGE IN

MUSCLE VARIABLES

TEMPORAL

CHANGE IN

OUTCOME

MEASURE

COFOUNDER

ADJUSTMENT RESULTS/EXPLAINED VARIANCE

stretching, cardiovascular

fitness, education, home

exercises

C: Sling exercises (40 minutes

each): maintaining neutral

spine position during different

preset sling exercises to

activate both deep and

superficial trunk muscles,

home exercises

(iv) was measured by

M-mode USI

Age, gender, BMI,

pain duration and

initial FABQ-P were

adjusted for.

Age, initial fear

avoidance for

physical activity,

gender, BMI,

and pain duration

were adjusted for.

and change in (i), (iii), and (iv); the

change variables (i, iii, and iv)

explained 47% of total variance

of pain at 1-year follow-up.

However, P value was not

reported.

In a logistic regression model,

with reference to patients with

low baseline (i) and negative

posttreatment temporal

change in (i), the OR for clinically

important pain reduction at 1-year

follow-up (NPRS $ 2)

for patients with:

–Low baseline and positive

change in (i) (OR = 14.7; 95%

CI = 2.41, 89.56)

–High baseline and negative

change in (i) (OR = 2.41; 95%

CI = .37, 15.58)

–High baseline and positive

change in (i) (OR = .20; 95%

CI = .01, 2.95)

Logistic regression of clinically

important reduction in pain

(NPRS $2) at 1-year

follow-up in relation to:

–Change in (i) OR = 1.14;

95% CI = .99, 1.30

–Change in (ii) OR = 1.04;

95% CI = .94, 1.14

–Change in (iv) OR =

.98; 95% CI = .95, 1.02

Vasseljen and

Fladmark,

2010102

(same cohort

as Vasseljen

et al, 2012,103

and Unsgaard-

Participants with chronic

nonspecific LBP recruited from

general practitioners or

physiotherapists and by

advertising among staff at a

local hospital (n = 109); Group

A (n = 36): 40.9 6 11.5 years;

A randomized controlled trial.

Each group received weekly

sessions over 8 weeks.

A: Motor control exercises

(40 minutes each): training

function of TrA, LM, and pelvic

floor muscle, cocontraction of

(i) TrA lateral slide

during ADIM

(ii) TrA-CR during ADIM

(iii) OI contraction

thickness ratio during ADIM

All measurements were

NPRS Confounders

adjustment was

not mentioned

Participants were split into equal

third based on changes in

posttreatment pain

at 8 weeks irrespective of

treatment allocation.

Multiple regression to

regress temporal change

1.e6

TheJournalofPain

Deep

TrunkMuscles

andBack

Pain—

ASystem

aticReview

Table 1. Continued

STUDYSTUDY POPULATION; AVERAGE AGE 6

STANDARD DEVIATION; FOLLOW-UP INTERVENTION

TEMPORAL CHANGE IN

MUSCLE VARIABLES

TEMPORAL

CHANGE IN

OUTCOME

MEASURE

OFOUNDER

DJUSTMENT RESULTS/EXPLAINED VARIANCE

Tondel et al,

2012100)

Group B (n = 36):

43.4 6 10.2 years; Group C

(n = 37): 36 6 10.3 years; 8-

week

deep trunk muscles and pelvic

floor muscles, incorporation of

deep trunk muscles contraction

during functional tasks, home

exercises

B: General exercises (1 hour):

general trunk strengthening,

stretching, cardiovascular

fitness, education, home

exercises

C: Sling exercises (40 minutes

each): maintaining neutral

spine position during different

preset sling exercises to

activate both deep and

superficial trunk muscles,

home exercises

measured with

B-mode USI

in NPRS on change in:

(i) b = .249; R2 = .0%; P = .67

(ii) b = 1.165; R2 = 10.0%; P < .01

Increased (ii) and reduced (iii)

together explained 18% of

variance in the multiple

regression model that regressed

temporal change in NPRS on

temporal change in muscle

variables score at 8-week

Vasseljen

et al, 2012103

(same cohort

as Unsgaard-

Tondel

et al, 2012,100

and Vasseljen

and Fladmark,

2010102)

Participants with chronic

nonspecific LBP recruited from

general practitioners or

physiotherapists and by

advertising among staff at a

local hospital (n = 109); Group

A (n = 36): 40.9 6 11.5 years;

Group B (n = 36):

43.4 6 10.2 years; Group C

(n = 37): 36 6 10.3 years; 8-

week

A randomized controlled trial.

Each group received weekly

sessions over 8 weeks.

A: Motor control exercises

(40 minutes each): training

function of TrA, LM, and

pelvic floor muscle, cocontraction

of deep trunk muscles and pelvic

floor muscles, incorporation of

deep trunk muscles contraction

during functional tasks,

home exercises

B: General exercises (1 hour):

general trunk strengthening,

stretching, cardiovascular fitness,

education, home exercises

C: Sling exercises (40 minutes

each): maintaining neutral spine

position during different preset

sling exercises to activate both

deep and superficial trunk

muscles, home exercises

(i) Anticipatory onset of

abdominal muscles

during contralateral

rapid arm flexion as

measured by M-mode USI

NPRS Age, g nder, BMI, pain

dur ion, and baseline

FAB -P were adjusted

for.

Results obtained based on pooled

data. Linear regression showed no

association between changes in

NPRS and change in (i) after the

8-week intervention. R2 < 2%

(P value was not reported)

With reference to patients with slow

baseline and no/slow change of

(i), the OR for posttreatment

clinically important pain

reduction (NPRS $ 2) in

patients with:

–Slow baseline and faster change

of (i) (OR = 1.03; 95%

CI = 1.00, 1.07).

–Fast baseline and faster change

of (i) (OR = 1.00; 95%

CI = .96, 1.05).

–Slow baseline and faster change

of (i) (OR = 1.04; 95% CI = .99,

1.09).

Wonget

alTheJournalofPain

1.e7

CA

e

at

Q

Table 1. Continued

STUDYSTUDY POPULATION; AVERAGE AGE 6

STANDARD DEVIATION; FOLLOW-UP INTERVENTION

TEMPORAL CHANGE IN

MUSCLE VARIABLES

TEMPORAL

CHANGE IN

OUTCOME

MEASURE

COFOUNDER

ADJUSTMENT RESULTS/EXPLAINED VARIANCE

A study that investigated dynamic morphometry of both TrA and LM

Koppenhaver

et al, 201163Patients with LBP and with or

without leg symptoms

recruited from the physical

therapy of an army medical

center and university campus

advertisements (n = 81);

34.4 6 12.6 years; 1 week

An uncontrolled study. Each

patient received 2 sessions of

spinal manipulative therapy 3–

4 days apart.

TrA percent thickness

change as measured

by B-mode USI during

ADIM and ASLR;

LM percent thickness

change at L4/5 or L5/S1

measured with B-mode

USI during CALT

Modified

ODI

Age, gender, and BMI

were adjusted for.

Alteration in TrA percent thickness

change was not presented in the

stepwise regression model for the

clinically important improvement

of modified ODI score. Age, sex,

and BMI was entered in the first

step. In step 2, the explanatory

variable had to have a significance

level of < .05 to enter the model

and significance level of > .1 to be

removed from the model.

Only 1-week change in contracted

LM thickness at L4/5 level

explained 7% of clinically

important improvement of

modified ODI score at 1-week;

b = .28; R2 = .07; P = .023

Studies that investigated dynamic morphometry or endurance of LM

Fritz

et al, 201131Patients with LBP and with or

without leg symptoms

recruited from physical therapy

clinics and community-based

advertisements (n = 51);

33.3 6 12.9 years; 1 week

An uncontrolled study. Patients

received 2 sessions of spinal

manipulative therapy 3–4

days apart.

LM percent thickness

change at L4/5 as

measured by B-mode

USI during CALT

Modified

ODI

No confounder

was adjusted for

Correlation between the

1-week percentage ODI

improvement and:

–3- to 4-day change in LM

recruitment, r = .006

(nonsignificant but

P value was not reported)

–1-week change in LM

recruitment,

r = �.096 (nonsignificant but P

value was not reported)

Mohseni-Bandpei

et al, 200679,*

Patients with chronic LBP from

general practitioners and

community physiotherapists

(n = 120); Group A (n = 60):

34.8 6 10.6 years; Group B

(n = 60): 37.2 6 10.2 years;

immediately after treatment �2

to 11 sessions (both clinical

outcomes and muscle testing)

A randomized controlled study.

A: Manipulation/exercise

group: spinal manipulation to

lumbar spine or sacroiliac

joint 1 individualized exercises

for an average of 4 sessions,

once or twice per week

B: Ultrasound/exercise group:

therapeutic ultrasound

Median frequency slope

of LM muscle during

Biering-Sorenson test as

measured by surface

EMG

ODI

VAS

No adjustment for

confounders

Results obtained from personal

communication using pooled

data:

–The posttreatment change of

median frequency slope was not

significantly correlated to

posttreatment changes of

(1) (r = .06, P = .72) or (2) (r = .07,

P = .67)

1.e8

TheJournalofPain

Deep

TrunkMuscles

andBack

Pain—

ASystem

aticReview

Table 1. Continued

STUDYSTUDY POPULATION; AVERAGE AGE 6

STANDARD DEVIATION; FOLLOW-UP INTERVENTION

TEMPORAL CHANGE IN

MUSCLE VARIABLES

TEMPORAL

CHANGE IN

OUTCOME

MEASURE

COFOUNDER

ADJUSTMENT RESULTS/EXPLAINED VARIANCE

treatment 1 individualized

exercises for an average of 6

sessions, once or twice per week.

Studies that investigated static morphometry of LM

Hides

et al, 199639Patients with acute LBP

(<3 weeks) were recruited from

accident and emergency

department in a hospital

(n = 41); Group A (n = 20):

30.9 6 6.5 years, Group B

(n = 21): 31 6 7.9 years; 4 and

10 weeks

A randomized controlled study

with 4-week treatment in

either group. A: Medical

management 1 motor

control exercise for LM and TrA

cocontraction in different

positions.

B: Medical management:

minor analgesics and

nonsteroidal anti-inflammatory

agents

LM CSA asymmetry from

L2 to S1 as measured

by B-mode USI

RMDQ

VAS

MPQ

No adjustment for

confounders

No significant correlation between

changes in RMDQ, VAS, or MPQ

scores and change in CSA

asymmetry of LM in Group B

patients (P valuewas not reported)

The correlation in Group

A was not reported.

Willemink

et al, 2012111,*

Patients with chronic nonspecific

LBP from a physiotherapy

practice in a hospital (n = 20);

46.2 6 9.7 years; 12-week; 24-

week

An uncontrolled study. Each

patient received isolated

lumbar extensors resistance

training using the Lower Back

Revival System. For the first

12 weeks, it was once per

week. In the following

12 weeks, the training

frequency depended on

patients’ availability.

FCSA of LM from L3

to S1 in MRI

1. RMDQ

2. PSFS score

3. Global

perceived

effect

No adjustment for

any confounders

Results obtained from personal

communication using pooled

data: Correlation between

temporal changes in FCSA and

temporal changes in:

From baseline to 12-week

1. r = �.07; CI = �.55, .44

2. r = .22; CI = �.31, .65

3. r = .06; CI = �.45, .54

From baseline to 24-week

1. r = �.48; CI = �.79, .02

2. r = �.58; CI = �.84, �.12

3. r = �.21; CI = �.64, .32

From 12-week to 24-week

1. r = �.32; CI = �.70, .21

2. r = �.13; CI = �.59, .39

3. r = �.31; CI = �.70, .22

Abbreviations: ADIM, abdominal drawing-in maneuver; TrA/OI, transversus abdominis/obliquus internus; EMG, electromyography; ODI, Oswestry Disability Index; VAS, visual analog scale; OI, obliquus internus; B-mode, brightness-mode;

RMDQ, Roland Morris disability questionnaire; PSFS, patient-specific functional scale; CPR, clinical prediction rule; NPRS, numerical pain rating scale; OE, obliquus externus; TDI, tissue Doppler imaging; LM, lumbar multifidus; M-

mode, motion-mode; FABQ-P, Fear-Avoidance Beliefs Questionnaire physical activity; BMI, body mass index; OR, odds ratio; TrA-CR, ratio of transversus abdominis muscle thickness at contraction to the resting thickness; ASLR, active straight

leg raise test; CALT, contralateral arm lifting task; CSA, total cross-sectional area; MPQ, McGill Pain questionnaire; FCSA, functional cross-sectional area; MRI, magnetic resonance imaging.

*The corresponding author gave the correlation data from personal communication.

Wonget

alTheJournalofPain

1.e9

1.e10 The Journal of Pain Deep Trunk Muscles and Back Pain—A Systematic Review

A comprehensive search strategy was employed usingmedical subheadings and keywords that encompassedthe themes of changes in multifidus or transversus ab-dominis characteristics, prediction, interventions, andclinical outcomes in patients with nonspecific LBP(see Appendix I for the detailed search strategy). Toidentify potential studies that assessed TrA/LM charac-teristics using different measurement methods, therewas no restriction on the type of muscle measurementtechniques. Additional searches were completed inClinicalTrial.gov, National Institutes of Health ClinicalCenter Clinical Research Studies Database, and CurrentControlled Trials Registers in order to identify ongoingstudies. The corresponding authors of the included ar-ticles and 10 prominent researchers who have pub-lished more than 5 articles in the relevant area werecontacted to identify additional publications (Fig 1).Last, the corresponding authors of potentially eligiblearticles (Appendix II and Table 1), which reported bothtemporal changes in TrA/LM characteristics (ie, posttestscores of a muscle variable at a given measurementtime point minus its corresponding baseline scores)and temporal changes in clinical outcomes (ie, posttestscores of a clinical outcome at a given measurementtime point minus its corresponding baseline scores)without analyzing the relation between the 2 do-mains, were contacted to seek relevant information(Fig 1).

Selection CriteriaThe detailed selection criteria for inclusion are shown

in Appendix III. A study was considered eligible for in-clusion if it was a randomized controlled trial or a non-randomized controlled trial investigating the relationbetween temporal changes in a muscle variable andthe corresponding changes in clinical outcomes of pa-tients treated for nonspecific LBP using nonsurgical(ie, conservative) methods. A study was also eligiblefor inclusion if it analyzed the relation between post-treatment absolute scores of muscle variable(s) andposttreatment absolute values of clinical outcome(s)with adjustment for the respective baseline values.Further, systematic reviews and meta-analyses thatwere relevant to our research questions were alsoeligible for inclusion. A muscle variable was defined asa feature related to the morphometry, histology, or acti-vation of TrA/LM that was hypothesized to be modifiedby a treatment and thought to be associated withchanges in clinical outcomes (Table 2). The definitionsof temporal change, static and dynamic morphometry,histology, muscle activation, clinical outcomes, andnonspecific LBP are listed in Table 3.

Study SelectionIn the first stage of screening, the identified studies

were stored in a Web-based citation management pro-gram (RefWorks, version 2.0; RefWorks-COS, Bethesda,MD) (Fig 1). Once duplicates were removed, 2 re-viewers (A.Y.L.W. and M.F.) used standardized formsto screen the titles and abstracts independently

according to the selection criteria. The selectionprocess was piloted on 100 citations and any disagree-ment was resolved via discussion to ensure the consis-tency between reviewers. Articles denoted as eligibleby either reviewer were included for the second-stage screening, in which the full-text copies of thestudies were screened using first-stage procedures. Ifconsensus could not be reached via discussion, a thirdreviewer (E.C.P.) arbitrated any persistent disagree-ment. The relevant citations in the reference lists ofthe included articles were searched. Forward citationtracking with Scopus and Web of Science was per-formed to identify relevant articles that had cited theincluded studies. The interreviewer reliability at eachscreening stage was examined by percent agreementand Kappa coefficients (k). Kappa was interpreted as.00 to .20 for poor agreement; .21 to .40 for fair agree-ment; .41 to .60 for moderate agreement; .61 to .80 forgood agreement; and .81 to 1.00 for almost perfectagreement.38,91

Risk of Bias AssessmentThe 2 reviewers assessed the risk of bias of the included

studies independently using an assessment toolmodifiedfrom checklists used in similar systematic reviews.5,108

The modifications aimed to enhance the risk of biasassessment of the study population and confounders(Table 4) using the recommendations on evaluatingprognostic studies.4,9,36 The maximum score of this toolis 19. The cut-off score for differentiating a high-quality study was 50% of maximum score.73 A highassessment score indicates low risk of bias. As the cut-off score was arbitrary, sensitivity analyses using 60%and 70% cut-off scores were used to examine the robust-ness of these cut-off scores.115

The risk of bias of the included systematic reviews, ifany, was evaluated by the Assessment of Multiple Sys-tematic Reviews (AMSTAR). AMSTAR is a reliable toolwith documented face, construct, and external validityfor assessing the risk of bias of systematic reviews.86,87

The interreviewer reliability of the risk of bias assess-ment was analyzed by percent agreement and intraclasscorrelation coefficient (ICC model 3.1) of the final score.The discrepancy in scores rated by the 2 reviewers wasresolved in a consensus meeting.

Data ExtractionTwo reviewers (A.Y.L.W. and M.F.) extracted the

following information independently from eachincluded study: 1) study design, 2) characteristics of par-ticipants, 3) sample size, 4) intervention (including type,dosage, duration, and frequency of a treatment), 5)muscle variables, 6) clinical outcomes, 7) statistical anal-ysis concerning the relation between muscle variablesand clinical outcomes, and 8) results at different assess-ment time points. If the included studies examined mul-tiple outcomes using different statistical methods, onlythe information that was relevant to our research ques-tions was extracted. A consensus meeting was held to

Table 2. Definitions of Various Muscle Variables in This Review and the Measurement MethodsUsed in the Included Studies

MUSCLES

MUSCLE VARIABLES

DEFINITIONS MEASUREMENTS METHODSMAIN CATEGORIES SUBCATEGORIES

TrA Thickness change during

contraction

TrA percent thickness change [(Contracted muscle thickness –

resting thickness)/resting

thickness] � 100%63

B-mode USI during ADIM, active

straight leg raise maneuver60,63

or an isometric knee flexion/

extension task25

TrA-CR Thickness atmaximum contraction/

resting thickness73; or (Thickness

at maximum contraction/resting

thickness) � 10100

M-mode USI73 during ADIM; or B-

mode USI during ADIM100,102,103

Lateral slide TrA lateral slide Maximum lateral displacement of

V-shaped midline border of TrA

from its resting to fully

contracted position102

B-mode USI during ADIM100,102,103

Anticipatory onset TrA onset Earliest point at which EMG signal

of TrA increased above baseline

level established over the

preceding 100-ms interval96

Fine-wire EMG during rapid arm

flexion test96,98

TrA/OI onset Earliest change in EMG signals of

TrA and OI muscles underneath

the surface electrodes77

Surface EMG during rapid arm

flexion test10,77

Lateral abdominal muscles onset The earliest change in thickness or

tissue velocity of lateral

abdominal muscles above

baseline level73,103

M-mode USI100,103 or TDI73 during

rapid arm flexion test

LM Thickness change during

contraction

LM percent thickness change [(Contracted muscle thickness –

resting thickness)/resting

thickness] � 100%63

B-mode USI during CALT31,63

Endurance LM MF slope MF = frequency that divides the

area of the EMG power

spectrum in half; MF slope is

calculated from the rate of

median frequency shift using

linear regression.79

Surface EMG during Biering-

Sorenson endurance test79

Static morphometry CSA asymmetry (CSA of the affected side at the

most affected vertebral level/

CSA of the unaffected side at the

same level) � 100%39

B-mode USI on resting LM39

FCSA CSA of the muscle isolated

from fat111MRI on resting LM111

Abbreviations: B-mode, brightness-mode; ADIM, abdominal drawing-in maneuver; TrA-CR, transversus abdominis contraction thickness ratio measured as the

transversus abdominis contracted thickness divided by transversus abdominis resting thickness; M-mode, motion-mode; EMG, electromyography; OI, obliquus in-

ternus; TDI, tissue Doppler imaging; LM, lumbar multifidus; CALT, contralateral arm lifting task; MF, median frequency; CSA, cross-sectional area; FCSA, functional

cross-sectional area; MRI, magnetic resonance imaging.

Wong et al The Journal of Pain 1.e11

ensure the accuracy and agreement of the extracteddata between the reviewers.

Data Synthesis and AnalysisIf the study design, population (such as age, gender,

or chronicity of LBP), type of muscle variable, outcomemeasure scales, treatments and clinical outcomes, andfollow-up time were homogeneous, meta-analysis wasconsidered for the studies with low risk of bias basedon the 50% cut-off score. Notably, meta-analysisfor each muscle variable was planned for differentfollow-up duration categories (ie, immediately aftertreatments, less than 1 year after treatments, and at

least 1 year after treatments). Meta-analysis withrespect to the secondary objectives (ie, gender; acute,subacute, and chronic LBP; age between 18 and 40 years,41 and 60 years, and above 60 years) was also planned.However, because our final results showed heteroge-neous data, meta-analysis was precluded and the infor-mation was summarized qualitatively.Two ad hoc subgroup analyses were conducted to

investigate the effects of treatment type and treatmentduration on the identified relations because theincluded studies used diverse treatments with differentdurations. Specifically, the treatment type subgroupanalysis qualitatively analyzed studies in 4 categories:1) spinal manipulative therapy, 2) motor control

Table 3. Definitions of Terms Used in ThisReview

TERMS DEFINITIONS

Temporal change/

alteration

Change in a muscle variable or a clinical

outcome as result of conservative

treatments between a given time point

and baseline (ie, posttest minus pretest

score)

Static morphometry The measurement of resting architecture

of a muscle (eg, shape, cross-sectional

area, thickness, and pennation angle)110

Dynamic morphometry The measurement of change in muscle

morphometry during contraction (eg,

transversus abdominis muscle thickness

change during contraction)110

Histology The examination of the microscopic

composition of a muscle11

Muscle activation The change in myoelectric signals of a

contracting muscle as measured by

electromyography (eg, an increase in

root mean square or median frequency

slope of electromyographic signals)12,20

Clinical outcomes Subjective and objective clinical measures

(eg, pain intensity, LBP-related disability

questionnaire scores, frequency of

recurrence, or number of LBP-related

sick leave)

Nonspecific LBP Pain of unknown cause located between

the 12th rib and buttocks, with or

without leg pain4,9,115

1.e12 The Journal of Pain Deep Trunk Muscles and Back Pain—A Systematic Review

exercises, 3) back-strengthening exercises, and 4) mixedtreatments (studies that used pooled data from multipletreatments). The treatment duration subgroup analysisqualitatively analyzed studies based on treatment dura-tion of #4 weeks and >4 weeks.The level of evidence (strong, moderate, limited, no,

and inconclusive evidence) of each variablewas classifiedaccording to the consistency of the research findings andthe risk of bias of the included studies (Table 5).14

If $ 75% of the relevant studies reported that a musclevariable demonstrated an association in the same direc-tion, the evidence was considered consistent.14,115

Statistical analyses were conducted using SPSSsoftware, version 18.0 (SPSS Inc, Chicago, IL).

Results

Literature SearchThe database searches resulted in 2,122 references

(Fig 1). Two thousand two articles were excluded basedon title, abstract, and duplication. One hundred twentyarticles were considered for full-text reading. Because14 full-text articles reported both temporal changes inTrA/LM characteristics and corresponding temporalchanges in clinical outcomes without evaluating theirassociations (Appendix II), the corresponding authorsof these articles were contacted. Three of these authorssubsequently provided unpublished information in theform of correlation coefficients between temporal

changes in TrA/LM characteristics and temporal changesin clinical outcomes.10,79,111 The others did not providesuch information because of various reasons (ie, invalidcontact [n = 1], no response from the authors [n = 6],loss of original data [n = 2], no intention to calculatethe correlation because of no temporal change inmuscle variables [n = 1], and no time to analyze thedata [n = 2]). As a result, only 3 additional articleswere included in the present review. Overall, 15articles10,25,31,39,60,63,73,77,79,96,98,100,102,103,111 from 12patient cohorts met the inclusion criteria. Thepercentage agreements between the 2 reviewers atthe first and second stages of screening were 84.8%(k = .70) and 93.3% (k = .78), respectively.

Risk of Bias AssessmentThe results of the risk of bias assessments are presented

in Table 6. Although 3 articles reported findings froman identical patient cohort, the risk of bias of thesearticles was evaluated separately because theyreported different muscle variables at different timepoints without duplicating information.100,102,103 Theassessment scores of the included studies ranged from5 to 13 out of a possible of 19 points. The rawagreement between the 2 reviewers’ scores was 73.3%(ICC3,1 = .92).When the 50% cut-off point for bias was used, 10 of 15

articles were classified as low risk of bias (ie, high quality)(Table 6).25,39,60,63,73,79,96,102,103,111 This number included2 of the 3 articles included after personal communicationwith the corresponding authors (Table 6).79,111 If the cut-off score was set at 60%, only 4 articles were categorizedas low risk of bias (Table 6).25,63,73,111 When 70% cut-offscore was used, all included articles were labeled as highrisk of bias (Table 6).Because the data obtained from the included studies

was clinically heterogeneous (ie, different clinicaloutcome measures and inconsistent follow-up duration)andmany of these studies did not report the actual associ-ation estimates,39,60,63,73,77,96,103 meta-analysis was pre-cluded.101,102 In fact, no more than 3 studies reporteddata with adequate homogeneity in terms of musclevariables, outcome variables, and measures ofassociation, which precluded the preparation ofmeaningful funnel plots.

Characteristics of the StudiesThe characteristics of patients, interventions, muscle

variables, and main findings of all the includedstudies are reported in Table 1. Elevenstudies10,25,60,63,73,77,96,98,100,102,103 analyzed thetreatment-related changes in TrA, whereas 5 arti-cles31,39,63,79,111 investigated LM (Table 1). Of these articles,1 analyzed changes in both TrA and LM.63 Seven studiesused prospective case series design,31,60,63,73,77,98,111

whereas 8 studies adopted randomized controlled trialdesign.10,25,39,79,96,100,102,103 Thirteen included articlesinvestigated temporal changes in dynamic morphometryor activation of TrA/LM.10,25,31,60,63,73,77,79,96,98,100,102,103

Two included studies investigated the temporal

Table 4. Risk of Bias Assessment Tool for Prospective Cohort Studies (Modified From Wesselset al108)

CRITERIA SCORE

Patient population

A Positive if an operational definition of cases, and inclusion and exclusion criteria were described. 1/�/?

B Positive if clear description of the source population (cases and controls should be from the same population) was described 1/�/?

C Positive if patients were representative of clinical practice 1/�/?

D Positive if an inception patient cohort (defined in relationship to onset of symptoms at the time of recruitment) was recruited 1/�/?

E Positive if clinical and demographic characteristics were described 1/�/?

Study attrition/follow-up

F Positive if the follow-up was at least 1 year* 1/�/?

G Positive if the dropout/loss to follow-up was <20% of the total sample 1/�/?

H Positive if the reasons for attrition and the descriptive characteristics of this group were given 1/�/?

Muscle variables

I Positive if a priori clear definition and rationale for potential treatment-targeted variable(s) were providedy 1/�/?

J Positive if the treatment-targeted variable(s) were measured using blinded, standardized, and validated measurements

(intraclass-correlation coefficients > .7 or k > .4 for intrarater and interrater reliability; if previous studies done by the same

research group have reported high reliability estimates for the measurement method(s), this criterion is also rated positively)

1/�/?

Intervention

K Positive if intervention was described in detail (including duration, frequency, dosage, and control for confounding activities) 1/�/?

Clinical outcome measures

L Positive if data on outcome are collected using standardized, blinded (except self-reported), and validated measurements

(intraclass-correlation coefficients > .7 or k > .4 for intrarater and interrater reliability; if previous studies done by the same

research group have reported high reliability estimates for the measurement method(s), this criterion is also rated positively)

1/�/?

Confounding variables

M Positive if important confounding variables were clearly defined and measured with valid and reliable measurements for all

potential relations (intraclass-correlation coefficients > .7 or k > .4 for intrarater and interrater reliability; if previous studies

done by the same research group have reported high reliability estimates for the measurement method(s), this criterion is also

rated positively)

1/�/?

Statistical analysis and data presentation

N Positive if mean changes in independent variables and outcomes through treatment were presented 1/�/?

O Positive if the data analysis included a stratified or multivariate analysis of association* 1/�/?

P Positive if the statistical model usedwas appropriate for the outcome studied and themeasures of association estimatedwith this

model were described (including confidence intervals)

1/�/?

Q Positive if the number of cases in the final multivariate model was at least 10 times the number of independent variables in the

analysisz1/�/?

R Positive if the analysis controlled for confounders 1/�/?

S Positive if for every single variable remaining in the model statistical measured were reported (either correlation coefficient, odds

ratio, Wilks’s Lambda, R2 or b) or if it was reported that the variable was not significant or explained no significant amount of

variance

1/�/?

*Both the muscle variables and the clinical outcomes should be measured simultaneously at the follow-up to be scored positive for this question. 1 = yes; � = no;

? = unclear.

yA priori clear definition and rationale for potential treatment process variable(s) should be indicated in the corresponding clinical trials registry.

zIf the model did not include a multivariate model, rate it as positive if the number of participants was at least 10.

Wong et al The Journal of Pain 1.e13

alterations in static morphometric changes in LM.39,111

There was no included article reporting the histology ofTrA/LM. Although this review planned to include variousclinical outcomes (such as return-to-work and recurrenceof LBP), the included studies assessed the relation betweentemporal alterations in only 3 clinical outcome variables(ie, pain intensity, disability, andglobal perceived recovery)and temporal changes in muscle variables. The includedstudies investigated 12 conservative interventions for pa-tients with acute (n = 1),39 subacute (n = 1),60 and chronicLBP (n = 13) (Table 1).10,25,31,63,73,77,79,96,98,100,102,103,111

Findings of the StudiesThe relation between temporal change in various TrA/

LM muscle variables over time and temporal change inclinical outcomes using a 50% cut-off of quality is sum-marized below.

Temporal Changes in TrA and Clinical Out-comes

Relation Between Temporal Alteration in TrA Thick-ness Change During Contraction and Temporal Changein Disability and Pain Intensity. Four included studiesexamined the relation between temporal changes inthis muscle variable (Table 2) and temporal change indisability using regression63 or correlation analyses(Table 1).25,60,73 Three high-quality studies found nosignificant correlation between these attributes(Table 1).60,63,73 Another high-quality study reportedconflicting results regarding the association betweentemporal alterations in TrA thickness change duringcontraction and temporal changes in 2 functional scores(Table 1).25

Three high-quality studies found no statistically signif-icant correlation between temporal alterations in TrA

Table 5. Level of Evidence

LEVEL OF EVIDENCE

Strong Consistent results ($75%) from at least 2 high-quality

studies

Moderate 1 high-quality study and consistent findings ($75%)

in 1 or more low-quality studies

Limited Findings in 1 high-quality cohort or consistent results

($75%) among low-quality studies

No No study identified

Conflicting Inconsistent results irrespective of study quality

Adapted from Cornelius et al.Adapted from Cornelius et al.14

1.e14 The Journal of Pain Deep Trunk Muscles and Back Pain—A Systematic Review

thickness change during contraction and temporalchange in pain intensity (Table 1).25 Another researchgroup reported that temporal increases in TrA contrac-tion thickness ratio explained 10% of the pre- to post-treatment pain reduction in chronic LBP patientsimmediately after 8-week exercise interventions(P < .01)102 but did not explain the absolute or clinicallyimportant pain reduction of the same cohort at 1-yearfollow-up (Table 1).100

In summary (Tables 1 and 7), there was strong evi-dence to support the absence of a relation betweentemporal alteration in TrA thickness change duringcontraction and temporal improvement in disability orpain intensity among patients with chronic LBP. Howev-er, there was only limited evidence in support of no rela-tion between immediate posttreatment temporalalteration in TrA thickness change during contractionand the long-term pain reduction in patients withchronic LBP.Relation Between Temporal Alteration in TrA Percent

Thickness Change During Contraction and Posttreat-mentGlobal Perceived Recovery. Onehigh-quality studyfound that temporal alteration in TrA percent thicknesschange was not correlated with posttreatment globalperceived recovery (Table 1).25 As such, there is limited

Table 6. Risk of Bias Assessment Scores of the Incl

ARTICLES

RISK OF BIAS ASSESSMENT

A B C D E F G H I J K

Hides et al, 199639 1 1 1 1 1 � 1 � ? 1 1

Marshall et al, 200677,111 1 1 ? � 1 � 1 � ? � 1

Mohseni-Bandpei et al, 200679 1 1 1 � 1 � 1 � ? � 1

Tsao and Hodges, 200898 1 � ? � � � � � ? ? 1

Tsao et al, 201096 1 � ? � 1 � 1 1 ? 1 1

Ferreira et al, 201025 1 1 1 � 1 � 1 1 � 1 1

Fritz et al, 201131 1 1 ? ? 1 � 1 � � 1 1

Koppenhaver et al, 201163 1 1 ? � 1 � 1 � ? 1 1

Konitzer et al, 201160 1 1 1 � 1 � 1 � ? 1 1

Brook et al, 201210 1 � ? � 1 � 1 � � ? 1

Mannion et al, 201273 1 1 1 � 1 � 1 � � ? 1

Willemink et al, 2012111 1 1 1 � ? � 1 1 � � 1

Vasseljen and Fladmark, 2010102 1 1 ? � 1 � � � � 1 1

Unsgaard-Tondel et al, 2012100 1 1 ? � 1 � � � � � 1

Vasseljen et al, 2012103 1 1 1 � 1 � � � � 1 1

*Cut-off score at 50%: 9.5; cut-off score at 60%: 11.4; cut-off score at 70%: 13.3.

evidence for the lack of a significant relation betweenthe 2 attributes (Table 7).Relation Between Temporal Change in Anticipatory

Onset of TrA/Lateral Abdominal Muscle Activation orMorphometric Change and Temporal Change inDisability or Pain Intensity. Two high-quality73,96 and 3low-quality studies10,77,98 found no significantcorrelation between temporal alterations in anticipatoryonset of TrA/lateral abdominal muscle activation/morphometric change and alterations in disability ofpatients with chronic LBP during treatment (Table 1).Similarly, 3 high-quality73,96,103 and 2 low-quality

studies77,98 reported no significant correlationbetween temporal changes in anticipatory onset ofTrA/lateral abdominal muscle activation/mor-phometric change and immediate posttreatmenttemporal pain reductions in patients with chronicLBP (Table 1). However, conflicting results were notedfor the relation between posttreatment temporalchange in this muscle variable (posttreatment scoreminus its baseline score) and clinically important painreduction of chronic LBP patients at 1-year follow-up(pain score at 1-year follow-up minus its baselinescore) (Table 1).100

Overall, there was strong evidence to support theabsence of a relation between temporal alterations inanticipatory onset of TrA/lateral abdominal muscles ac-tivity and temporal changes in disability or pain intensityin patients with chronic LBP. However, therewas conflict-ing evidence for the relation between posttreatmentchange in this muscle variable and change in pain at 1-year follow-up (Tables 1 and 7).Relation Between Temporal Change in TrA Lateral

Slide and Temporal Change in Pain Intensity. Oneresearch group found that change in TrA lateral slideover the course of treatments was not significantly corre-lated with temporal change in absolute/clinically impor-tant pain intensity (numerical pain rating scale $ 2) in

uded Studies

CRITERIA

TOTAL

RISK OF BIAS WITH CUT-OFF AT*

L M N O P Q R S 50% 60% 70%

1 � 1 � � 1 � � 11 Low High High

� 1 1 1 � � � 1 9 High High High

1 � 1 � � 1 � 1 10 Low High High

1 � 1 � � � � 1 5 High High High

1 � 1 � � 1 � 1 10 Low High High

1 � 1 � 1 1 � 1 13 Low Low High

1 � � � � 1 � 1 9 High High High

1 1 1 1 � 1 1 1 13 Low Low High

1 � 1 � � 1 � 1 11 Low High High

1 � 1 � 1 1 � 1 9 High High High

1 � 1 1 1 1 � 1 12 Low Low High

1 � 1 � 1 1 � 1 11 Low Low High

1 ? 1 1 � 1 � 1 10 Low High High

1 1 � 1 1 � 1 � 9 High High High

1 1 � � � � 1 1 10 Low High High

Table 7. Level of Evidence for Morphometry, Histology or Activation of Transversus Abdominis or Lumbar Multifidus as a Treatment-TargetedVariable With Cut-Off Point of Bias at 50%

CHANGES IN MUSCLE

VARIABLES SUBCATEGORIES

CHANGES IN

CLINICAL OUTCOMES

NUMBER OF COHORT STUDIES WITH

POSITIVE OR NO RESULTS AND RISK

OF BIAS OF THE STUDIES (HIGH/LOW)

RELATIONS

LEVEL OF EVIDENCE

POSITIVE/NEGATIVE

RELATION NO RESULTS

INDIVIDUALOUTCOME MEASURE

OVERALL AT

50% CUT-OFF

OVERALL AT

60% CUT-OFF

OVERALL AT

70% CUT-OFFHIGH LOW HIGH LOW

Disability

TrA thickness change during

contraction

TrA percent thickness change RMDQ 25,* Positive Limited Strong Strong Limited

Modified ODI 63,*,60,*,y No Strong

PSFS 25,* No Limited

TrA-CR RMDQ 73,* No Limited

TrA anticipatory onset TrA onset PSFS 98 96,*,y No Moderate Strong Limited Limited

TrA/OI onset ODI 10,77 No Limited

Lateral abdominal

muscles onset

RMDQ 73,* No Limited

LM LM percent

thickness change

Modified ODI 63,* 31 Inconclusive Conflicting Conflicting Conflicting Conflicting

MF slope ODI 79,*,y No Limited Limited Limited Limited

CSA asymmetry RMDQ 39,*,y No Limited Limited Limited Limited

FCSA as measured by MRI RMDQ 111,* No Limited Conflicting Conflicting Conflicting

FCSA as measured by MRI PSFS 111,* 111,* Inconclusive Conflicting

Pain

TrA thickness change during

contraction

TrA percent thickness

change

VAS/NPRS 25,*,60,*,y No Strong Strong Strong Limited

TrA-CR PGRS 73,* No Limited

NPRS (short term) 102,*,y Positive Limited

NPRS (1-year) 100 No Limited

Clinical important NPRS

reduction (1-year)

100 No Limited

TrA Lateral slide TrA lateral slide NPRS (short term) 102,*,y No Limited Limited Limited Limited

NPRS (long term) 100 No Limited

Clinical important NPRS

reduction (1-year)

100 No Limited

Low baseline TrA lateral

slide and increased TrA

lateral slide compared

with those with stable

or reduced slide

NPRS 100 Positive Limited Limited Limited Limited

Wonget

alTheJournalofPain

1.e15

Table 7. Continued

CHANGES IN MUSCLE

VARIABLES SUBCATEGORIES

CHANGES IN

CLINICAL OUTCOMES

NUMBER OF COHORT STUDIES WITH

POSITIVE OR NO RESULTS AND RISK

OF BIAS OF THE STUDIES (HIGH/LOW)

RELATIONS

LEVEL OF EVIDENCE

POSITIVE/NEGATIVE

RELATION NO RESULTS

INDIVIDUALOUTCOME MEASURE

OVERALL AT

50% CUT-OFF

OVERALL AT

60% CUT-OFF

OVERALL AT

70% CUT-OFFHIGH LOW HIGH LOW

TrA anticipatory onset TrA onset VAS 98 95,*,y No Limited Strong Limited Limited

TrA/OI onset Current or last week VAS 10 No Limited

Lateral abdominal

muscles onset

PGRS (short-term)

(long-term)

73,* No Limited

NPRS short term 103,*,y No Limited

NPRS (1-year) 100 Positive Limited

Clinically important NPRS

reduction (1-year)

100 No Limited

LM MF slope VAS 79,*,y No Limited Limited Limited Limited

CSA asymmetry MPQ 39,*,y No Limited Limited Limited Limited

VAS 39,*,y No Limited

Global impression of recovery

TrA TrA percent

thickness change

11-point global

impression of recovery

25,* No Limited Limited Limited Limited

LM FCSA 7-point global

perceived effect

111,* No Limited Limited Limited Limited

Abbreviations: high, cohort study with high risk of bias; low, cohort study with low risk of bias; RMDQ, Roland Morris disability questionnaire; ODI, Oswestry disability index; inconclusive, contradictive findings regarding the factor as a

treatment modifier or prognostic factor; conflicting, conflicting level of evidence; PSFS, patient-specific functional scale; limited, limited level of evidence; strong, strong level of evidence; no, result did not support the factor as a treatment

modifier or a prognostic factor; positive, larger temporal increase of muscle variable was associated with larger temporal decrease in disability/pain reduction; TrA-CR, transversus abdominis contraction ratio measured as the transversus

abdominis contracted thickness divided by transversus abdominis resting thickness; TrA/OI, transversus abdominis/obliquus internus; LM, lumbar multifidus; MF, median frequency; CSA, cross-sectional area; FCSA, functional cross-sectional

area; MRI, magnetic resonance imaging; VAS, visual analog scale; NPRS, numeric pain rating scale; PGRS, pain graphic rating scale; MPQ, McGill pain questionnaire.

*The article was rated as a high risk of bias paper at 70% cut-off score.

yThe article was rated as a high risk of bias paper at 60% cut-off score.

1.e16

TheJournalofPain

Deep

TrunkMuscles

andBack

Pain—

ASystem

aticReview

Wong et al The Journal of Pain 1.e17

the short and long terms (Table 1).100,102 However, theynoted that patients with low baseline TrA lateral slideand an increased posttreatment TrA lateral slide had ahigher odds ratio for clinically important painreduction than those with low baseline but stable orreduced posttreatment TrA lateral slide at 1-yearfollow-up (Table 1).100

In summary (Tables 1 and 7), there was limited evi-dence that temporal change in TrA lateral slidewas unre-lated to the longitudinal changes in pain intensity inboth the short and long terms. For patients with lowbaseline TrA lateral slide, there was limited evidence tosupport the relation between posttreatment temporalincrease in TrA lateral slide and the pain reduction at1-year follow-up.

Temporal Changes in LM and Clinical Out-comes

Relation Between Temporal Alteration in LM PercentThickness Change During Contraction and TemporalChange in Disability or Pain Intensity. One high-qualitystudy reported that a temporal increase in LM percentthickness change during contraction explained 7% ofthe variance in posttreatment disability reduction(Table 1).63 However, no significant correlation wasnoted between these attributes in a low-quality study(Table 1).31 None of the included articles investigated therelation between temporal change in this muscle vari-able and temporal pain reduction.In conclusion, there was conflicting evidence for the

relation between temporal alteration in LM percentthickness change during contraction and temporalchange in disability. There was no evidence regardingthe relation between temporal change in this musclevariable and temporal change in pain intensity(Tables 1 and 7).Relation Between Temporal Change in LM Endurance

and Temporal Change in Pain or Disability. From theresults of a single study (Tables 1 and 7),79 there waslimited evidence to support the absence of a relationbetween temporal changes in LM endurance (ie, medianfrequency slope) during the Biering-Sorenson endurancetest and temporal changes in disability/pain intensity.Relation Between Temporal Change in Static

Morphometry of LM and Temporal Change in Disabilityor Pain Intensity. One high-quality study revealed thattemporal changes in LM cross-sectional area asymmetrywere not correlated with temporal changes in disabilityor pain intensity in patients receiving pharmaceuticaltreatments (Table 1).39 Another high-quality studyshowed inconsistent findings regarding the relation be-tween temporal changes in LM functional cross-sectionalarea and temporal changes in disability (unpublisheddata from corresponding author) (Table 1).111

In short (Tables 1 and 7), limited evidence supportedthat temporal change in LM cross-sectional area asym-metry was unrelated to temporal change in pain inten-sity or disability. However, conflicting evidence wasnoted for the relation between temporal changes in LMfunctional cross-sectional area and temporal disabilityimprovements over various time treatment intervals.

Relation Between Temporal Alteration in LM Func-tional Cross-Sectional Area and Posttreatment GlobalPerceived Recovery. A high-quality study found no sig-nificant correlation for these attributes (Table 1).111 As aresult, limited evidence supported a lack of a relationbetween temporal change in LM functional cross-sectional area and global perceived recovery followingconservative treatments (Table 7).

Temporal Changes in Histology of TrA/LM andClinical Outcomes

Because no identified research investigated the rela-tion between temporal changes in histology of TrA/LMand temporal changes in clinical outcomes, there wasno evidence to substantiate this relation.Effects of Gender, LBP Chronicity, and Age: The Iden-

tified Relations. Because none of the included articlesreported results stratified by gender, LBP chronicity, orage, subgroup analyses related to the secondary objec-tive of this review were not possible.Effects of Treatment Type or Treatment Duration on

the Identified Relations. There was consistently nosignificant association between temporal changesin various muscle variables and temporal changes indisability/pain intensity for all the treatment typesin the subgroup analysis (spinal manipulative ther-apy,31,60,63 motor control exercises,73,96,98 back-strengthening exercises,77,111 and most of the mixedtreatment category; Table 8).10,39,79,100,102,103 However,conflicting evidence was noted for the associations (ie,positive or no) between temporal alterations in TrAthickness change during contraction and temporalchanges in disability/pain intensity for the mixedtreatment category (Table 8).25,100,102

Treatment duration had no effect on altering theabove-mentioned relations except for the conflictingreport on the association between temporal alterationin TrA thickness change during contraction and tempo-ral change in disability (Table 8).25,73 The reason forsuch difference was the inconsistent results from astudy25 in the >4-week category that reported thattemporal increase in TrA thickness change duringcontraction was related to temporal reduction in theRoland Morris disability scores but unrelated to tem-poral change in the patient-specific functional scores(Tables 1 and 8).25

Effect of Cut-Off Scores on Evidence Synthesis. If thecut-off score of the risk of bias assessment was set at 60%or 70% of themaximum score, the level of evidence for 4identified associations would be affected (Table 7). Spe-cifically, if a 60% or 70% cut-off point was used, the ev-idence for the absence of a relation between temporalchanges in anticipatory onset of TrA activity and tem-poral changes in disability or pain intensity would bealtered from strong to limited (Table 7). Additionally, if a70% cut-off pointwas used, the evidence for the lack of arelation between temporal alterations in TrA thicknesschange during contraction and temporal changes indisability or pain intensity would be changed fromstrong to limited (Table 7).

Table 8. Subgroup Analyses and Level of Evidence Based on Treatment Types and Treatment Durations

MUSCLE

TEMPORAL CHANGES IN

MUSCLE VARIABLES

SUBCATEGORIES

CHANGES IN

CLINICAL OUTCOMES

LEVEL OF EVIDENCE

ALL STUDIESCOMBINED

TREATMENT TYPE SUBGROUP TREATMENT DURATION SUBGROUPS

SPINALMANIPULATIVE THERAPY

MOTOR

CONTROL EXERCISES MIXED STRENGTHENING EXERCISES #4 WEEK >4 WEEK

TrA TrA thickness change

during contraction

Disability No (strong) No (limited) No (limited) 1 and no

(conflicting)

N.A. No (limited) 1 and no (conflicting)

Pain No (strong) No (limited) No (limited) 1 and no

(conflicting)

N.A. No (limited) No (strong)

Perceived global

recovery

N.A. N.A. N.A. No (limited) N.A. N.A. No (limited)

TrA/lateral abdominal

muscle onset

Disability No (strong) N.A. No (strong) No (limited) No (limited) No (limited) No (strong)

Pain No (strong) N.A. No (strong) No (limited) N.A. No (limited) No (strong)

TrA lateral slide Pain No (limited) N.A. N.A. No (limited) N.A. N.A. No (limited)

LM LM percent thickness

change

Disability 1 and no (conflicting) 1 and no (conflicting) N.A. N.A. N.A. 1 and no

(conflicting)

N.A.

LM endurance Disability No (limited) N.A. N.A. No (limited) N.A. No (limited) N.A.

Pain No (limited) N.A. N.A. No (limited) N.A. No (limited) N.A.

LM CSA asymmetry Pain No (limited) N.A. N.A. No (limited) N.A. No (limited) N.A.

LM FCSA Disability � and no (conflicting) N.A. N.A. N.A. � and no (conflicting) N.A. � and no (conflicting)

Perceived global

recovery

No (limited) N.A. N.A. N.A. No (limited) N.A. No (limited)

Abbreviations: mixed, using pooled data from various treatments for analysis; CSA, cross-sectional area; FCSA, functional cross-sectional area; no, no relation; N.A., not applicable; limited, limited level of evidence;1, larger temporal increase

of muscle variable was associated with larger temporal decrease in disability/pain reduction; �, larger temporal increase of muscle variable was associated with smaller temporal decrease in disability/pain reduction; conflicting, conflicting

level of evidence.

1.e18

TheJournalofPain

Deep

TrunkMuscles

andBack

Pain—

ASystem

aticReview

Wong et al The Journal of Pain 1.e19

DiscussionTo our knowledge, this is the first systematic review to

summarize evidence regarding whether changes inmorphometry, histology, or activation of TrA/LMfollowing conservative interventions explained changesin clinical outcomes of nonspecific LBP. Our results under-scored the absence of an association for TrA and anuncertain association for LM. The conflicting findingsfor these LMmuscle variablesmight be due to the limitednumber of relevant publications.The emphasis of restoring TrA/LM muscle variables in

patients with LBP has prevailed for 2 decades.Although individual reports have shown that post-treatment improvement of LBP symptoms/disabilitymay not correlate with posttreatment improvementof TrA/LM muscle variables,39,73,96 many scientistssuggested that patients in LBP remission havingabnormal TrA/LM muscle variables might be at higherrisk for LBP recurrence.17,42,70 This premise arises inpart from prospective research that showed thatpatients with acute LBP displayed LM asymmetry.39,41

Although all patients in that study were pain freeafter receiving 4 weeks of medical or motor controlexercise treatment, only patients who received motorcontrol exercise restored normal LM symmetry at 4-and 10-week follow-ups.39 Interestingly, patientsreceiving motor control exercise had lower recurrencerates of LBP at 1- and 3-year follow-ups, but the asso-ciation between change in muscle asymmetry andrecurrence was not quantified.41 Despite the fact thatthe researchers did not adjust for confounders (eg,psychological factors or occupation) that could haveaffected LBP recurrence, the premise of restoringdeep trunk muscle morphometry/function to reduceLBP recurrence is well accepted. Our findings substan-tiate that there is no, or low, correlation between post-treatment temporal changes in TrA/LM and temporalchanges in clinical outcomes, although our resultscannot support or disprove the premise that restora-tion of TrA/LM reduces LBP recurrence, as no includedstudy analyzed the correlation between these 2domains.

Comparison With Prior WorkOur findings also concur with systematic reviews

that reported no significant relation between tempo-ral changes in other physical parameters (eg, backmuscle strength or spinal stiffness) and temporalchanges in clinical outcomes.67,89,108 Although theresults may be ascribed to various factors (eg, poorassociation between physical capacity/deconditioningand LBP development,35,88 complex processing ofself-perceived pain/disability,44,98 heterogeneousmuscle activation strategies in patients,49 and diverseunderlying causes of nonspecific LBP7), it may alsobe attributable to the heterogeneity of methodolo-gies and treatments. Therefore, the characteristics ofthe included articles are discussed below in order toidentify the research gaps and future researchdirections.

Characteristics of the Included Studies

Quality of the Included Studies

Seven of 15 included studies were not primarilydesigned to analyze the associations between differentmuscle variables and clinical improve-ments.10,31,39,79,98,103,111 Therefore, their risk of biasassessment scores might be suboptimal. However, thisproblem was unlikely because our findings showedthat 12 of 15 included articles demonstrated low risk ofbias at the 50% cut-off point, and the sensitivity analysesat different cut-off points only yielded slightly differentconclusions.

Participants

The participants in all included studies, except 1,39

could be classified as chronic LBP patients with low tomoderate pain intensity and disability. Because the prog-nosis of patients with chronic LBP involves complex inter-actions among physical and psychosocial factors,29,37 ourfindingsmay not be generalizable to patientswith acute/subacute/recurrent LBP.

Physical Tests

Eight25,31,39,60,63,100,102,103 of the 15 included studiesused B-mode or M-mode USI to measure TrA/LMdynamic morphometry.61,62,114 However, dynamicmorphometry might be affected by multiple factors(eg, resting state of the muscle, the competing tensionsfrom surrounding tissues, the out-of-planemuscle move-ment, and the operator’s experience109,115). Thesefactors might introduce variability and attenuatecorrelation. To minimize these potential confoundingeffects, future studies may consider usingelectromyography to directly monitor TrA/LMactivity.47,96

Although LBP has known to cause localized deficits inLM characteristics (eg, LM cross-sectional area asymme-try),19,43,59 many included studies31,63,79,111 did notselect the location of LM examination based on thepain location. As such, they might fail to detect thetemporal improvements in the affected LM musclevariables.

Treatment Types

Because motor control exercises may restore themorphometry39,42 and central motor control of TrA/LMin patients with LBP,96-98 patients receiving suchtreatment are expected to have greater improvementsin TrA/LM impairments and a stronger relationbetween temporal changes in TrA/LM characteristicsand temporal changes in clinical outcomes. However,the 3 included studies73,96,98 that solely investigatedmotor control exercises did not support this relation(Tables 1 and 8). These studies73,96,98 investigatedfeedforward activation of TrA/lateral abdominalmuscles during rapid arm movements. However, thevalidity of such a test for quantifying lumbarneuromuscular deficits has been questioned3,80,103

given the high variability of TrA feedforward activation

1.e20 The Journal of Pain Deep Trunk Muscles and Back Pain—A Systematic Review

in individuals with3,10,33,95,103 and without LBP.74,102

Therefore, their findings22,71,73,96,98,104 should beinterpreted with caution.Ten10,25,39,73,77,96,98,100,102,103 included articles

prescribed motor control exercises (Table 1), but only 4small studies showed significant temporal changes inTrA onset timing,96,98 TrA contraction thickness,73 or LMstatic morphometry.39 Regarding the other treatments,only 2 showed significant temporal changes in LM endur-ance or contraction thickness.63,79 Collectively, theabsence of temporal muscle changes might explain thenonsignificant associations in this review.

Treatment and Follow-Up Duration

The treatment and follow-up duration in mostincluded studies was <8 weeks. Although our subgroupanalysis revealed that treatment duration had minimaleffect on the overarching conclusions of this review(Table 8), stronger association may be observed if partic-ipants continue the treatments or home exercises forlonger duration. Additionally, some studies only treatedpatients once or twice.31,60,63 One may argue that suchtreatment intensity might not elicit significant changesin TrA/LM static morphometry though it may causechanges in motor control.97 Taken together, futurestudies should investigate the optimal treatment type/in-tensity/duration for restoring the deficits of various mus-cle variables.One included study reassessed participants’ clinical

outcomes100 but not the TrA dynamic morphometry at1-year follow-up. Therefore, their findings might notreveal the true relations between both attributes overtime. To investigate whether long-term changes in TrA/LM characteristics are related to long-term clinical out-comes/recurrence of LBP,20,53,65 future studies shouldcollect data from both attributes at the follow-ups.

Confounders

Psychosocial factors (eg, catastrophization, depres-sion, and fear avoidance behavior) may influence theprognosis or satisfaction of patients with LBP.2,30,38

However, only 1 research group adjusted forpsychological factors in their statistical analyses.100,103

Analyzing the relation between temporal changes inTrA/LM features and temporal clinical improvementswithout accounting for psychosocial confounders maybias the results.94,99

Statistical Analysis

Only 6 of 15 articles reported the relation betweentemporal changes in TrA/LM dynamic morphometry/acti-vations and longitudinal clinical improvements usingmultiple regression,63,73,77,100,102,103 which analyzes therelative importance of each of the multiple explanatoryvariables in explaining the variance of the dependentvariable (eg, LBP/disability).32 Because LBP clinicaloutcomes are influenced by multiple factors, a multivar-iate analysis would be more appropriate. Intriguingly, 3of the 6 studies using multivariate analyses found signif-icant associations.63,100,102

Most included studies (Table 1) might be underpow-ered to detect significant associations given their smallsample sizes. For instance, the theoretical sample sizefor a correlation of medium effect size is 85 (Pearson cor-relation coefficient = .3, a = .05, power = .8, 2-tailedtest).83 However, 11 of 15 included studies did not meetthis requirement.10,25,31,39,60,63,73,77,96,98,111 Further,many of these studies did not report the actualassociation estimates,39,60,63,73,77,96,103 which affect thereader’s ability in appraising the clinical importance ofthe association and the likelihood of the type II error.

Strengths of This ReviewThis review had several strengths. Comprehensive

search strategies were adopted to optimize the identifi-cation of candidate publications in 4 languages. Further,we contacted the corresponding authors of the includedand potential articles and prominent researchers to iden-tify potential articles for this review. Our review protocolwas registered in PROSPERO to allow checking for selec-tive reporting.Our riskof bias assessment toolwasmodified froma tool

used in a similar systematic review108 using the recommen-dations for evaluating prognostic studies.9,36 Further,because our assessment tool has not been validated, weconducted sensitivity analyses using 60% and 70% cut-off scores to test the robustness of our assessments.115

LimitationsBecause gray literature was excluded and funnel plots

were not constructed owing to the heterogeneity of theincluded studies,4,21 this review may be subject topublication bias that tends to publish positive studies.However, as our findings revealed that the relationsbetween the 2 attributes were absent or weak, the riskof publication bias was estimated to be low. Inaddition, our conclusions might have been stronger ordifferent had all authors provided additional datawhen contacted by our study team.

ImplicationsAlthough this review reveals the absence of an asso-

ciation between temporal changes in various TrA/LMcharacteristics and temporal changes in clinical out-comes, multiple factors may account for the negativefindings. Future large-scale prospective studies shouldbe conducted to unveil the potential relation betweentemporal changes in these attributes by 1) investi-gating treatments separately with longer follow-up toallow more change to be detected on both the musclevariables and clinical outcomes, 2) adjusting forconfounders, and 3) targeting patients with definitebaseline TrA/LM deficits. Finally, because temporalchanges in TrA/LM characteristics may mediate therelation between other physical parameters (eg, spinalstiffness) and clinical outcomes,31 future studies shouldadopt novel technologies to quantify these interrela-tions.17,50,113,114

Wong et al The Journal of Pain 1.e21

AcknowledgmentsThe authors would like to thank librarian Linda Seale

for the design of literature search. We would like toexpress our gratitude to Dr. Tasha Stanton, a postdoc-toral research fellow from the School of Health Sciencesat the University of South Australia, for her professional

input. The authors would also like to acknowledge thetimely responses of the authors of studies included inthis review. Last, we thank Lydia Dani for assisting inthe procedures of removing the duplicated citationsand the identifiable information of the potentialcitations.

References

1. Akbari A, Khorashadizadeh S, Gholam A: The effect ofmotor control exercise versus general exercise on lumbarlocal stabilizing muscles thickness: Randomized controlledtrial of patients with chronic low back pain. J Back Musculo-skeletal Rehabil 21:105-112, 2008

2. Al-Obaidi SM, Beattie P, Al-Zoabi B, Al-Wekeel S: The rela-tionship of anticipated pain and fear avoidance beliefs tooutcome in patients with chronic low back pain who are notreceiving workers’ compensation. Spine 30:1051-1057, 2005

3. Allison GT, Morris SL, Lay B: Feedforward responses oftransversus abdominis are directionally specific and actasymmetrically: Implications for core stability theories. J Or-thop Sports Phys Ther 38:228-237, 2008

4. Altman DG: Systematic reviews of evaluations of prog-nostic variables. BMJ 323:224-228, 2001

5. Ari€ens GAM, van Mechelen W, Bongers PM, Bouter LM,van der Wal G: Psychosocial risk factors for neck pain: A sys-tematic review. Am J Ind Med 39:180-193, 2001

6. Bajek S, Bobinac D, Bajek G, Vrani�c TS, Lah B,Dragojevi�c DM: Muscle fiber type distribution in multifidusmuscle in cases of lumbar disc herniation. Acta MedOkayama 54:235-241, 2000

7. Balagu�e F, Mannion AF, Pellis�e F, Cedraschi C: Non-specificlow back pain. Lancet 379:482-491, 2012

8. Barker PJ, Guggenheimer KT, Grkovic I, Briggs CA,Jones DC, Thomas CDL, Hodges PW: Effects of tensioningthe lumbar fasciae on segmental stiffness during flexionand extension: Young Investigator Award winner. Spine31:397-405, 2006

9. Beattie PF, Nelson RM: Evaluating research studies thataddress prognosis for patients receiving physical therapycare: A clinical update. Phys Ther 87:1527-1535, 2007

10. Brooks C, Kennedy S, Marshall PWM: Specific trunk andgeneral exercise elicit similar changes in anticipatory posturaladjustments in patients with chronic low back pain: A ran-domized controlled trial. Spine 37:E1543-E1550, 2012

11. Brown S, Gregory D, Carr J, Ward S, Masuda K, Lieber R:Adaptation to the multifidus muscle in response to experi-mentally induced intervertebral disc degeneration. Spine36:1728-1736, 2011

12. Brown SHM,McGill SM: A comparison of ultrasound andelectromyography measures of force and activation toexamine the mechanics of abdominal wall contraction.Clin Biomech (Bristol, Avon) 25:115-123, 2010

13. Coghlan S, Crowe L, McCarthypersson U, Minogue C,Caulfield B: Neuromuscular electrical stimulation trainingresults in enhanced activation of spinal stabilizing mus-cles during spinal loading and improvements in pain rat-ings. Conf Proc IEEE Eng Med Biol Soc 2011:7622-7625,2011

14. Cornelius LR, van der Klink JJL, Groothoff JW, Brouwer S:Prognostic factors of long termdisability due tomental disor-ders: A systematic review. J Occup Rehabil 21:259-274, 2011

15. Coste JL, Lefran ois GR, Guillemin F, Pouchot J, FrenchStudy Group for Quality of Life in Rheumatology: Prognosisand quality of life in patients with acute low back pain: In-sights from a comprehensive inception cohort study.Arthritis Rheum 51:168-176, 2004

16. Critchley D, Coutts F: Abdominal muscle function inchronic low back pain patients measurement with real-time ultrasound scanning. Physiotherapy 88:322-332, 2002

17. D’hooge R, Cagnie B, Crombez G, Vanderstraeten G,Achten E, Danneels L: Lumbar muscle dysfunction duringremission of unilateral recurrent nonspecific low-backpain: Evaluation with muscle functional MRI. Clin J Pain29:187-194, 2013

18. Danneels LA, Vanderstraeten GG, Cambier DC,Witvrouw EE, De Cuyper HJ: CT imaging of trunk musclesin chronic low back pain patients and healthy control sub-jects. Eur Spine J 9:266-272, 2000

19. Dickx N, Cagnie B, Parlevliet T, Lavens A, Danneels L: Theeffect of unilateral muscle pain on recruitment of the lum-bar multifidus during automatic contraction. An experi-mental pain study. Man Ther 15:364-369, 2010

20. Dunn KM, Jordan K, Croft PR: Characterizing the courseof low back pain: A latent class analysis. Am J Epidemiol 163:754-761, 2006

21. Easterbrook PJ, Gopalan R, Berlin JA,Matthews DR: Pub-lication bias in clinical research. Lancet 337:867-872, 1991

22. EkmanM, Johnell O, Lidgren L: The economic cost of lowback pain in Sweden in 2001. Acta Orthop 76:275-284, 2005

23. Franca FR, Burke TN, Hanada ES,Marques AP: Segmentalstabilization and muscular strengthening in chronic lowback pain—A comparative study. Clinics 65:1013-1017, 2010

24. Ferreira PH, FerreiraML, Hodges PW: Changes in recruit-ment of the abdominal muscles in people with low backpain: Ultrasound measurement of muscle activity. Spine 29:2560-2566, 2004

25. Ferreira PH, Ferreira ML, Maher CG, Refshauge K,Herbert RD, Hodges PW: Changes in recruitment of transver-sus abdominis correlate with disability in people withchronic low back pain. Br J Sports Med 44:1166-1172, 2010

26. Fortin M, Macedo LG: Multifidus and paraspinal musclegroup cross-sectional areas of patients with low back painand control patients: A systematic review with a focus onblinding. Phys Ther 93:873-888, 2013

27. Franca FR, Burke TN, Caffaro RR, Ramos LA,Marques AP:Effects of muscular stretching and segmental stabilizationon functional disability and pain in patients with chroniclow back pain: A randomized controlled trial. J Manipula-tive Physiol Ther 35:279-285, 2012

1.e22 The Journal of Pain Deep Trunk Muscles and Back Pain—A Systematic Review

28. FreemanMD,WoodhamMA,WoodhamAW: The role ofthe lumbar multifidus in chronic low back pain: A review.PM&R 2:124-146, 2010

29. Friedrich M, Hahne J, Wepner F: A controlled examina-tion of medical and psychosocial factors associated withlowback pain in combinationwithwidespreadmusculoskel-etal pain. Phys Ther 89:786-803, 2009

30. Fritz JM, George SZ, Delitto A: The role of fear-avoidance beliefs in acute low back pain: Relationshipswith current and future disability and work status. Pain 94:7-15, 2001

31. Fritz JM, Koppenhaver SL, Kawchuk GN, Teyhen DS,Hebert JJ, Childs JD: Preliminary investigation of the mecha-nisms underlying the effects ofmanipulation: Exploration ofa multi-variate model including spinal stiffness, multifidusrecruitment, and clinical findings. Spine 36:1772-1781, 2011

32. Glass GV, Hopkins KD: Statistical Methods in Educationand Psychology, 3rd ed. Needham Heights, MA, Allyn & Ba-con, 1996

33. Gubler D, Mannion AF, Schenk P, Gorelick M,Helbling D, Gerber H, Toma V, Sprott H: Ultrasound tissueDoppler imaging reveals no delay in abdominal musclefeed-forward activity during rapid arm movements in pa-tients with chronic low back pain. Spine 35:1506-1513,2010

34. Haldeman S, Kopansky-Giles D, Hurwitz EL, Hoy D, MarkErwinW, Dagenais S, Kawchuk G, Stromqvist B,Walsh N: Ad-vancements in the management of spine disorders. BestPract Res Clin Rheumatol 26:263-280, 2012

35. Hamberg-van Reenen HH, Ari€ens GAM, Blatter BM, vanMechelen W, Bongers PM: A systematic review of the rela-tion between physical capacity and future low back andneck/shoulder pain. Pain 130:93-107, 2007

36. Hayden JA, Cot�e P, Bombardier C: Evaluation of thequality of prognosis studies in systematic reviews. AnnIntern Med 144:427-437, 2006

37. Hayden JA, Dunn KM, van der Windt DA, Shaw WS:What is the prognosis of back pain? Best Pract Res Clin Rheu-matol 24:167-179, 2010

38. Henschke N, Wouda L, Maher CG, Hush JM, vanTulderMW: Determinants of patient satisfaction 1 year afterpresenting to primary care with acute low back pain. Clin JPain 29:512-517, 2013

39. Hides J, Richardson C, Jull G: Multifidus muscle recoveryis not automatic after resolution of acute, first-episode lowback pain. Spine 21:2763-2769, 1996

40. Hides J, Wilson S, Stanton W, McMahon S, Keto H,McMahon K, Bryant M, Richardson C: An MRI investigationinto the function of the transversus abdominis muscle dur-ing ‘‘drawing-In’’ of the abdominal wall. Spine 31:E175-E178, 2006

41. Hides JA, Jull GA, Richardson CA: Long-term effects ofspecific stabilizing exercises for first-episode low backpain. Spine 26:E243-E248, 2001

42. Hides JA, Stanton W, McMahon S, Sims K, Richardson C:Effect of stabilization training on multifidus muscle cross-sectional area among young elite cricketers with low backpain. J Orthop Sports Phys Ther 38:101-108, 2008

43. Hides JA, Stokes MJ, Saide M, Jull GA, Cooper DH: Evi-dence of lumbar multifidus muscle wasting ipsilateral to

symptoms in patients with acute/subacute low back pain.Spine 19:165-172, 1994

44. Hodges PW: Pain andmotor control: From the laboratoryto rehabilitation. J Electromyogr Kinesiol 21:220-228, 2011

45. Hodges PW, Moseley GL, Gabrielsson A, Gandevia SC:Experimental muscle pain changes feedforward postural re-sponses of the trunk muscles. Exp Brain Res 151:262-271,2003

46. Hodges PW, Richardson CA: Altered trunk musclerecruitment in people with low back pain with upper limbmovement at different speeds. Arch Phys Med Rehabil 80:1005-1012, 1999

47. Hodges PW, RichardsonCA: Delayed postural contractionof transversus abdominis in low back pain associated withmovement of the lower limb. J Spinal Disord 11:46-56, 1998

48. Hodges PW, Richardson CA: Inefficient muscular stabili-zation of the lumbar spine associated with low back pain: Amotor control evaluation of transversus abdominis. Spine21:2640-2650, 1996

49. Hodges PW, Tucker K: Moving differently in pain: A newtheory to explain the adaptation to pain. Pain 152:S90-S98,2011

50. Hu Y, Wong YL, Lu WW, Kawchuk GN: Creation of anasymmetrical gradient of back muscle activity and spinalstiffness during asymmetrical hip extension. Clin Biomech(Bristol, Avon) 24:799-806, 2009

51. Hyun JK, Lee JY, Lee SJ, Jeon JY: Asymmetric atrophy ofmultifidus muscle in patients with unilateral lumbosacralradiculopathy. Spine 32:E598-E602, 2007

52. Ihlebæk C, Hansson TH, Lærum E, Brage S, Eriksen HR,Holm SH, Svendsrød R, Indahl A: Prevalence of low backpain and sickness absence: A ‘‘borderline’’ study in Norwayand Sweden. Scand J Public Health 34:555-558, 2006

53. Jellema P, van der Windt DAWM, van der Horst HE,Stalman WAB, Bouter LM: Prediction of an unfavourablecourse of low back pain in general practice: Comparison offour instruments. Br J Gen Pract 57:15-22, 2007

54. Jemmett RS, Macdonald DA, Agur AMR: Anatomical re-lationships between selected segmental muscles of the lum-bar spine in the context of multi-planar segmental motion:A preliminary investigation. Man Ther 9:203-210, 2004

55. J€oud A, Petersson IF, Englund M: Low back pain—Epide-miology of consultations. Arthritis Care Res 64:1084-1088,2012

56. Kalichman L, Hodges P, Li L, Guermazi A, Hunter DJ:Changes in paraspinal muscles and their association withlow back pain and spinal degeneration: CT study. Eur SpineJ 19:1136-1144, 2010

57. Kankaanpaa M, Taimela S, Airaksinen O, Hanninen O:The efficacy of active rehabilitation in chronic low backpain. Effect on pain intensity, self-experienced disability,and lumbar fatigability. Spine 24:1034-1042, 1999

58. Kiesel KB, Uhl T, Underwood FB, Nitz AJ: Rehabilitativeultrasound measurement of select trunk muscle activationduring induced pain. Man Ther 13:132-138, 2008

59. Kiesel KB, Underwood FB, Mattacola CG, Matacolla C,Nitz AJ, Malone TR: A comparison of select trunk musclethickness change between subjects with low back pain clas-sified in the treatment-based classification system and

Wong et al The Journal of Pain 1.e23

asymptomatic controls. J Orthop Sports Phys Ther 37:596-607, 2007

60. Konitzer LN, Gill NW, Koppenhaver SL: Investigation ofabdominal muscle thickness changes after spinal manipula-tion in patients who meet a clinical prediction rule for lum-bar stabilization. J Orthop Sports Phys Ther 41:666-674,2011

61. Koppenhaver S, Hebert J, Fritz J, Parent E, Teyhen D,Magel J: Reliability of rehabilitative ultrasound imaging ofthe transversus abdominis and lumbar multifidus muscles.Arch Phys Med Rehabil 90:87-94, 2009

62. Koppenhaver S, Hebert J, Parent E, Fritz J: Rehabilitativeultrasound imaging is a valid measure of trunk muscle sizeand activation during most isometric sub-maximal contrac-tions: A systematic review. Aust J Physiother 55:153-169,2009

63. Koppenhaver SL, Fritz JM, Hebert JJ, Kawchuk GN,Childs JD, Parent EC, Gill NW, Teyhen DS: Association be-tween changes in abdominal and lumbar multifidus musclethickness and clinical improvement after spinal manipula-tion. J Orthop Sports Phys Ther 41:389-399, 2011

64. Korff Von M, Deyo RA, Cherkin D, Barlow W: Back painin primary care: Outcomes at 1 year. Spine 18:855-862, 1993

65. Korff Von M, Miglioretti DL: A prognostic approach todefining chronic pain. Pain 117:304-313, 2005

66. Kumar S, Sharma VP, Shukla R, Dev R: Comparative effi-cacy of twomultimodal treatments onmale and female sub-groups with low back pain (part II). J Back MusculoskeletRehabil 23:1-9, 2010

67. Laird RA, Kent P, Keating JL: Modifying patterns ofmovement in people with low back pain—Does it help? Asystematic review. BMC Musculoskelet Disord 13:169, 2012

68. Lee W, Lee Y, Gong W: The effect of lumbar strength-ening exercise on pain and the cross-sectional area changeof lumbar muscles. J Phys Ther Sci 23:209-212, 2011

69. Luoto S, Heliovaara M, Hurri H, Alaranta H: Static backendurance and the risk of low-back pain. Clin Biomech (Bris-tol, Avon) 10:323-324, 1995

70. MacDonald D, Moseley GL, Hodges PW: Why do somepatients keep hurting their back? Evidence of ongoingback muscle dysfunction during remission from recurrentback pain. Pain 142:183-188, 2009

71. Manchikanti KN, Pampati V, Falco FJE, Hirsch JA: Growthof spinal interventional pain management techniques:Analysis of utilization trends and medicare expenditures2000 to 2008. Spine 38:157-168, 2013

72. Mannion A, K€aser L, Weber E, Rhyner A, Dvorak J,Muntener M: Influence of age and duration of symptomson fibre type distribution and size of the back muscles inchronic low back pain patients. Eur Spine J 9:273-281, 2009

73. MannionAF, Caporaso F, Pulkovski N, Sprott H: Spine sta-bilisation exercises in the treatment of chronic low backpain: A good clinical outcome is not associated withimproved abdominal muscle function. Eur Spine J 21:1301-1310, 2012

74. MannionAF, Pulkovski N, Schenk P, Hodges PW,GerberH,Loupas T, Gorelick M, Sprott H: A newmethod for the nonin-vasive determination of abdominal muscle feedforwardactivity based on tissue velocity information from tissueDoppler imaging. J Appl Phys 104:1192-1201, 2008

75. Marshall P, Murphy B: Delayed abdominal muscle onsetsand self-report measures of pain and disability in chroniclow back pain. J Electromyogr Kinesiol 20:833-839, 2010

76. Marshall PW, Murphy BA: Muscle activation changes af-ter exercise rehabilitation for chronic low back pain. ArchPhys Med Rehabil 89:1305-1313, 2008

77. Marshall PWM, Murphy BA: Evaluation of functionaland neuromuscular changes after exercise rehabilitationfor low back pain using a Swiss ball: A pilot study. J Manip-ulative Physiol Ther 29:550-560, 2006

78. MoherD, Liberati A, Tetzlaff J, AltmanDG, PRISMAGroup:Preferred reporting items for systematic reviews and meta-analyses: The PRISMA statement. Int J Surg 151:264-270, 2009

79. Mohseni-Bandpei MA, Critchley J, Staunton T,Richardson B: A prospective randomised controlled trial ofspinal manipulation and ultrasound in the treatment ofchronic low back pain. Physiotherapy 92:34-42, 2006

80. Morris SL, Lay B, Allison GT: Corset hypothesisrebutted—Transversus abdominis does not co-contract inunison prior to rapid armmovements. Clin Biomech (Bristol,Avon) 27:240-254, 2012

81. Parkkola R, Ryt€okoski U, Kormano M: Magnetic reso-nance imaging of the discs and trunk muscles in patientswith chronic low back pain and healthy control subjects.Spine 18:830-836, 1993

82. Pengel LHM: Acute low back pain: Systematic review ofits prognosis. BMJ 327:323, 2003

83. Portney L, Watkins M: Foundations of Clinical Research:Applications to Practice, 3rd ed. Upper Saddle River, NJ,Prentice Hall, 2009

84. Richardson C, Hodges P, Hides J: Therapeutic Exercise forLumbopelvic Stabilization: A Motor Control Approach forthe Treatment and Prevention of Low Back Pain, 2nd ed.Philadelphia, PA, Churchill Livingstone, 2004

85. Richardson C, Jull G, Hodges P, Hides J: Therapeutic Exer-cises for Spinal Segmental Stabilization in Low Back Pain:Scientific Basis and Clinical Approach. Philadelphia, PA,Churchill Livingstone, 1998

86. Shea BJ, Grimshaw JM, Wells GA, Boers M, Andersson N,Hamel C, Porter AC, Tugwell P, Moher D, Bouter LM: Devel-opment of AMSTAR: Ameasurement tool to assess themeth-odological quality of systematic reviews. BMC Med ResMethodol 7:10, 2007

87. Shea BJ, Hamel C, Wells GA, Bouter LM, Kristjansson E,Grimshaw J, Henry DA, Boers M: AMSTAR is a reliable andvalid measurement tool to assess the methodological qualityof systematic reviews. J Clin Epidemiol 62:1013-1020, 2009

88. Smeets RJEM, Wade D, Hidding A, Van Leeuwen PJCM,Vlaeyen JWS, Knottnerus JA: The association of physical de-conditioning and chronic low back pain: A hypothesis-oriented systematic review. Disabil Rehabil 28:673-693, 2006

89. Steiger F, Wirth B, de Bruin ED, Mannion AF: Is a positiveclinical outcome after exercise therapy for chronic non-specific low back pain contingent upon a correspondingimprovement in the targeted aspect(s) of performance? Asystematic review. Eur Spine J 21:575-598, 2012

90. Stroup DF, Berlin JA, Morton SC, Olkin I, Williamson GD,Rennie D, Moher D, Becker BJ, Sipe TA, Thacker SB: Meta-analysis of observational studies in epidemiology: A pro-posal for reporting. J Am Med Assoc 283:2008-2012, 2000

1.e24 The Journal of Pain Deep Trunk Muscles and Back Pain—A Systematic Review

91. Sun X, Briel M, Busse JW, You JJ, Akl EA, Mejza F,Bala MM, Mertz D, Diaz-Granados N, Vandvik PO,Malaga G, Srinathan SK, Dahm P, Johnston BC, Alonso-Coello P, Hassouneh B, Walter SD, Heels-Ansdell D,Bhatnagar N, Altman DG, Guyatt GH: Credibility of claimsof subgroup effects in randomised controlled trials: System-atic review. BMJ 344:e1553, 2012

92. Sung PS:Multifidimusclesmedian frequency before andafter spinal stabilization exercises. Arch Phys Med Rehabil84:1313-1318, 2003

93. Teyhen DS, Williamson JN, Carlson NH, Suttles ST,O’Laughlin SJ, Whittaker JL, Goffar SL, Childs JD: Ultrasoundcharacteristics of the deep abdominal muscles during theactive straight leg raise test. Arch Phys Med Rehabil 90:761-767, 2009

94. Tosteson TD, Buzas JS, Demidenko E, Karagas M: Powerand sample size calculations for generalized regressionmodels with covariate measurement error. Stat Med 22:1069-1082, 2003

95. Tsao H, Druitt TR, Schollum TM, Hodges PW: Motortraining of the lumbar paraspinal muscles induces immedi-ate changes in motor coordination in patients with recur-rent low back pain. J Pain 11:1120-1128, 2010

96. Tsao H, Galea MP, Hodges PW: Driving plasticity in themotor cortex in recurrent low back pain. Eur J Pain 14:832-939, 2010

97. Tsao H, Hodges PW: Immediate changes in feedforwardpostural adjustments following voluntary motor training.Exp Brain Res 181:537-546, 2007

98. Tsao H, Hodges PW: Persistence of improvements inpostural strategies following motor control training in peo-plewith recurrent low back pain. J Electromyogr Kinesiol 18:559-567, 2008

99. Turner EL, Perel P, Clayton T, Edwards P, Hern�andez AV,Roberts I, Shakur H, Steyerberg EW, CRASH trial collabora-tors: Covariate adjustment increased power in randomizedcontrolled trials: An example in traumatic brain injury. JClin Epidemiol 65:474-481, 2012

100. Unsgaard-Tøndel M, Nilsen TIL, Magnussen J,Vasseljen O: Is activation of transversus abdominis and obli-quus internus abdominis associated with long-term changesin chronic low back pain? A prospective study with 1-yearfollow-up. BMJ Sports Med 46:729-734, 2012

101. van Tulder MW, Assendelft WJJ, Koes BW, Bouter LM,Group TEBOTCCBR: Method guidelines for systematic re-views in the Cochrane Collaboration Back Review Groupfor Spinal Disorders. Spine 22:2323-2330, 1997

102. Vasseljen O, Fladmark AM: Abdominal muscle contrac-tion thickness and function after specific and general exer-cises: A randomized controlled trial in chronic low backpain patients. Man Ther 15:482-489, 2010

103. Vasseljen O, Unsgaard-Tøndel M, Westad C, Mork PJ:Effect of core stability exercises on feed-forward activationof deep abdominal muscles in chronic low back pain: A ran-domized controlled trial. Spine 37:1101-1108, 2012

104. Vos T, Flaxman AD, Naghavi M, Lozano R, Michaud C,Ezzati M, Shibuya K, Salomon JA, Abadalla S, Aboyans V,Abraham J, Ackeman I, Aggarwal R, Ahn SY, Ali MK,Alvarado M, Anderson HR, Anderson LM, Andrews KG,Atkinson C, Baddour LM, Bahalim AN, Barker-Collo S,Barrero LH, Bartels DH, Basanez MG, Baxter A, Bell ML,Benjamin EJ, Bennett D, Bernabe E, Bhalla K, Bhandari B,Bikbov B, Abdulhak AB, B G, Black JA, Blencowe H, Blore JD,

B F, Bolliger I, Bonaventure A, Boufous S, Bourne R,Bossussinesq M, Braithwaite T, Brayne C, Bridgett L,Brooker S, Brooks P, Brugha TS, Bryan-Hancock C, Bucello C,Buchbinder R, Buckle G, Budke CM, Burch M, Burney P,Burstein R, Calabria B, Campbell B, Canter CE, Carabin H,Carapetis J, Camona L, Cella C, Charlson F, Chen H,Cheung AT, Chou D, Chugh SS, Coffeng LE, Colan SD,Colquhoun S, Colson KE, Candon J, Connor MD, Cooper LT,Corriere M, Cortinovis M, Couville de Vaccaro K, Couser W,Cowie BC, Criqui MH, Cross M, Dabhadkar KC, Dahiya M,Dahodwala N, Damsere-Derry J, Danaei G, Davis A, DeLeo D, Degenhardt L, Dellavalle R, Delossantos A,Denenberg J, Derrett S, Jarlais Des DC, Dharmaratne SD,Dherani M, Diaz-Tome C, Dolk H, Dorsey ER, Driscoll T,Duber H, Ebel B, Edmond K, Elbaz A, Ali SE, Erskine H,Erwin PJ, Espindola P, Ewoigbokhan SE, Farazadfar F,Feigin V, Felson DT, Ferrari A, Ferri CP, Fevre EM,Finucane MM, Flaxman S, Flood L, Foreman K,Forouzanfar MH, Fowkes GR, Franklin R, Fransen M,Freeman MK, Gabbe BJ, Gabriel SE, Gakidou E, Ganatra HA,Garcia B, Gaspari F, Gillum RF, Gmel G, Gosselin R,Grainger R, Groeger J, Guillemin F, Gunnell D, Gupta R,Haagsma J, Hagan H, Halasa YA, Hall W, Haring D, Haro JM,Harrison JE, Havmoeller R, Hay RJ, Higashi H, Hill C, Hoen B,Hoffman H, Hotez PJ, Hoy D, Huang JJ, Ibeanusi SE,Jacobsen KH, James SL, Javis D, Jarsrasaria R, Jayaraman S,Johns N, Jonas JB, Karthikeyan G, Kassebaum N,Kawakami N, Keren A, Khoo JP, King CH, Knowlton LM,Kobusingye O, Koranteng A, Krishnamurthi R, Lalloo R,Laslett LL, Lathlean T, Leasher JL, Lee YY, Leigh J, Lim SS,LimbE, Lin JK, LipnickM, Lipshultz SE, LiuW, LoaneM,LockettOhno S, Lyons R, Ma J, Mabweijano J, Maclntyre MF,Malekzadeh R, Mallinger L, Manivannan S, Marcenes W,March L, Margolis DJ, Marks GB, Marks R, Matsumori A,Matzopoulos R, Mayosi BM, McAnulty JH, McDermott MM,McGill N, McGrath J, Medina-Mora ME, Meltzer M,Mensah GA, Merriman TR, Meyer AC, Miglioli V, Miller M,Miller TR, Mitchell PB, Mocumbi AO, Moffitt TE,Mokdad AA, Monsata L, Montico M, Moradi-Lakeh M,Moran A, Morawska L, Mori R, Murdoch ME, Mwaniki MK,Naidoo K, Nair MN, Naldi L, Narayan KMV, Nelson PK,Nelson RG, Nevitt M, Newton CR, Nolte S, Norman P,Norman R, O’Donnell M, O’Hanlon S, Olives C, Omer SB,Ortblad K, Osborne R, Ozgediz D, Page A, Pahari B,Pandian JD, Rivero AP, Patten SB, Pearce N, Padilla RP, Perez-Ruiz F, Perico N, Pesudovs K, Phillips D, Phillips MR, Pierce K,Pion S, Polanczyk GV, Polinder S, Pope CA III, Popova S,Porrini E, Pourmalek F, Prince M, Pullan RL, Ramaiah KD,Ranganathan D, Razavi H, Regan M, Rehm JT, Rein DB,RemuzziG, RichardsonK, Rivara FP, Roberts T, RobinsonC, Ro-driguez De Leon F, Ronfani L, Room R, Rosenfeld LC,Rushton L, Sacco RL, Saha S, Sampson U, Sanchez-Riera L,Sanman E, Schwebel DC, Scott JG, Segui-Gomez M,Shahraz S, Shepard DS, Shin H, Shivakoti R, Singh D,Singh GM, Singh JA, Singleton J, Sleet DA, Sliwa K, Smith E,Smith JL, Stapelberg NJC, Steer A, Steiner T, Stalk WA,Stovner LJ, Sudfeld C, Syed S, Tamburlini G, Tavakkoli M,Taylor HR, Taylor JA, Taylor WJ, Thomas B, Thomson WM,Thurston GD, Tleyjeh IM, Tonelli M, Towbin JA, Truelsen T,Tsilimbaris MK, Ubeda C, Undurranga EA, van der Werf MJ,van Os J, Vavilala MS, Venketasubramanian N, Wang M,Wang W, Watt K, Weatherall DJ, Weinstock MA,Weintraub R, Weisskopf MG, Weissman MM, White RA,Whiteford H, Wiersma ST, Wilkinson JD, Williams HC,Williams SRM, Witt E, Wolfe F, Woolf AD, Wulf S, Yeh PH,Zaidi AKM, Zheng ZJ, Zonies D, Lopez AD, Murray CJL: Yearslived with disability (YLDs) for 1160 sequelae of 289 diseasesand injuries 1990–2010: A systematic analysis for the GlobalBurden of Disease Study 2010. Lancet 380:2163-2196, 2012

105. Waddell G: 1987 Volvo award in clinical sciences: A newclinical model for the treatment of low-back pain. Spine 12:632-644, 1987

Wong et al The Journal of Pain 1.e25

106. Walker BF: The prevalence of low back pain: A system-atic review of the literature from 1966 to 1988. J Spinal Dis-ord 13:205-217, 2000

107. Wallwork TL, StantonWR, FrekeM,Hides JA: The effectof chronic low back pain on size and contraction of the lum-bar multifidus muscle. Man Ther 14:496-500, 2009

108. Wessels T, van Tulder M, Sigl T, Ewert T, Limm H,Stucki G: What predicts outcome in non-operative treat-ments of chronic low back pain? A systematic review. EurSpine J 15:1633-1644, 2006

109. Whittaker JL, Stokes M: Ultrasound imaging and mus-cle function. J Orthop Sports Phys Ther 41:572-580, 2011

110. Whittaker JL, Teyhen DS, Elliott JM, Cook K,Langevin HM, Dahl HH, Stokes M: Rehabilitative ultrasoundimaging: Understanding the technology and its applica-tions. J Orthop Sports Phys Ther 37:434-449, 2007

111. Willemink MJ, van Es HW, Helmhout PH, Diederik AL,Kelder JC, van Heesewijk JPM: The effects of dynamic isolatedlumbar extensor training on lumbar multifidus functionalcross-sectional area and functional status of patients withchronicnonspecific lowbackpain.Spine37:E1651-E1658,2012

112. Williams JM, Haq I, Lee RY: An investigation into theonset, pattern, and effects of pain relief on lumbarextensor electromyography in people with acute andchronic low back pain. J Manipulative Physiol Ther 36:91-100, 2013

113. Wong AY, Kawchuk G, Parent EC, Prasad N: Repeat-ability of instrumented spinal stiffness measurements usinga mechanical indentation device in individuals with andwithout low back pain. Man Ther 18:395-402, 2013

114. Wong AYL, Parent E, Kawchuk G: Reliability of 2 ultra-sonic imaging analysis methods in quantifying lumbar mul-tifidus thickness. J Orthop Sports Phys Ther 43:251-262,2013

115. Wong AYL, Parent EC, Funabashi M, Stanton TR,Kawchuk GN: Do various baseline characteristics of transver-sus abdominis and lumbar multifidus predict clinical out-comes in non-specific low back pain? A systematic review.Pain, 2013 [Epub ahead of print]

116. Zielinski KA, Henry SM, Ouellette-Morton RH,DeSamoMJ: Lumbarmultifidusmuscle thickness doesnotpre-dict patients with low back pain who improve with trunk sta-bilizationexercises.ArchPhysMedRehabil 94:1132-1138, 2013

Appendix I. Keywords for Literature Search andSearch Strategies

RELATED ITEMS KEY WORDS

Population Explode low back pain (MeSH), lower

back (MeSH), LBP, lumbar pain (MeSH),

lumbago (MeSH), lumbalgia (MeSH),

backache (MeSH), explode sciatica (MeSH)

Lumbar multifidus and

transversus abdominis

Lumbar multifidus (MeSH), SM, LM,

transversus abdominis (MeSH), TrA,

lumbar muscle (MeSH), back muscle

(MeSH), muscle activity, core muscle,

erector spinae (MeSH), stabilizing musc*

(MeSH), muscular (MeSH)

Study designs Explode cohort studies (MeSH), incidence

(MeSH), explode mortality (MeSH),

follow-up studies (MeSH), mortality

(MeSH), prognos*, predict*, course,

explode randomized control*, RCT, course

MEDLINE Search Strategies:1. (lumbo* or lumbar).mp. [mp=title, abstract, orig-

inal title, nameof substanceword, subject headingword, keyword heading word, protocol supple-mentary concept, rare disease supplementaryconcept, unique identifier]

2. exp Back/ or back.mp. or exp Lumbar Vertebrae/ orexp Lumbosacral Region/ or exp Coccyx/ or coc-cyx.mp. or torso.mp. or exp Torso/

3. 1 or 24. muscle, skeletal.mp. or exp Muscle, Skeletal/ or

muscul*.mp. or exp Muscular Atrophy/5. 3 and 46. (paraspinal or multifid* or back muscle or back ex-

tensor*).mp.7. (LM or MF).mp. [mp=title, abstract, original title,

name of substance word, subject heading word,keyword heading word, protocol supplementaryconcept, rare disease supplementary concept,unique identifier]

8. 3 and 79. 5 or 6 or 810. exp Abdominal Muscles/ or transversus abdomi-

nis.mp. or transverse abdominis.mp. or rectusabdominis.mp. or exp Rectus Abdominis/ orTrA.mp. or external oblique*.mp. or internal ob-lique*.mp.

11. (core or (stabilization or stabilizing)).mp.12. 5 and 1113. 10 or 1214. (change* or improv* or alter*).mp.15. 9 and 1416. back pain.mp. or exp Back Pain/ or exp Spinal

Diseases/ or lumbalgia.mp. or exp Low BackPain/ or LBP.mp. or sciatica.mp. or exp Sciatica/or (backache or lumbago or dorsalgia or lumbarpain).mp.

17. prognosis.mp. or exp Prognosis/ or incidence.mp.or exp Incidence/ or morbidity.mp. or exp

Morbidity/ or exp Risk Factors/ or (predict* orcourse or prognos*).mp.

18. exp Follow-Up Studies/ or followup.mp. or follow-up.mp. or exp Cohort Studies/ or cohort.mp. or expLongitudinal Studies/ or longitudinal.mp. or expProspective Studies/ or prospective.mp. or expRetrospective Studies/ or retrospective.mp.

19. 17 and 1820. 15 and 16 and 1921. limit 20 to (‘‘all adult (19 plus years)’’ and (Chinese

or English or French or Portuguese))22. 13 and 1423. 16 and 19 and 2224. limit 23 to (‘‘all adult (19 plus years)’’ and (Chinese

or English or French or Portuguese))

EMBASE Search Strategies1. lumbar.mp. or exp lumbar vertebra/ or exp lumbar

spine/ or exp back/ or exp lumbosacral spine/ orlumbo*.mp. or lumbosacral.mp. or ilium.mp. orexp iliac bone/ or coccyx.mp. or exp coccygealbone/ or exp trunk/ or trunk.mp. or torso.mp.

2. muscle atrophy/ or muscle contraction/ ormuscle.mp. or exp muscle/ or exp muscle function/or muscul*.mp. or exp musculoskeletal system/

3. 1 and 24. paraspinal.mp. or exp back muscle/ or back exten-

sor*.mp. or multifid*.mp.5. (LM or MF).mp.6. 2 and 57. 3 or 4 or 68. exp rectus abdominis muscle/ or exp abdominal

wall musculature/ or transverse abdominis.mp. ortransversus abdominis.mp. or rectus abdomi-nis.mp. or TrA.mp. or external oblique*.mp. or in-ternal oblique*.mp.

9. (core or (stabilization or stabilizing)).mp. or expspine stabilization/

10. 2 and 911. 8 or 1012. (change* or improv* or alter*).mp.13. 7 and 1214. back pain.mp. or exp backache/ or lumbago.mp.

or exp low back pain/ or LBP.mp. or (dorsalgiaor lumbalgia).mp. or sciatica.mp. or exp ischial-gia/

15. (follow-up or followup).mp. or exp follow up/ orexp longitudinal study/ or longitudinal.mp. orexp cohort analysis/ or cohort.mp. or prospecti-ve.mp. or exp prospective study/ or exp retrospec-tive study/ or retrospective.mp.

16. morbidity.mp. or exp morbidity/ or risk factor.mp.or exp risk factor/ or exp prediction/ or pre-dict*.mp. or progno*.mp. or course.mp. or expprognosis/or exp incidence/ or incidence.mp.

17. 15 and 1618. 13 and 14 and 1719. limit 18 to (abstracts and human and (Chinese or

English or French or Portuguese) and (adult <18to 64 years > or aged <651 years>))

1.e26 The Journal of Pain Deep Trunk Muscles and Back Pain—A Systematic Review

20. 11 and 1221. 14 and 17 and 2022. limit 21 to (abstracts and human and (Chinese or

English or French or Portuguese) and (adult <18to 64 years > or aged <651 years>))

CINAHL Search Strategies1. MM ‘‘Lumbar Vertebrae’’ OR Lumbar OR back OR

MM ‘‘Back’’ OR trunk OR MM ‘‘Torso’’ OR spineOR MM ‘‘Spine1’’ OR MM ‘‘Coccyx’’ OR lumbosa-cral OR MM ‘‘Lumbosacral Plexus1’’

2. MM ‘‘Musculoskeletal System1’’ OR MM ‘‘Mus-cles1’’ OR MM ‘‘Muscle, Skeletal1’’ OR muscle*

3. S1 and S24. (paraspinal) and (S2)5. (MM ‘‘Multifidus Muscles’’) OR ‘‘multifid*’’6. MF’’ OR back extensor* OR ‘‘LM’’7. S3 or S4 or S5 or S68. ((MM ‘‘Rectus Abdominis Muscles’’) OR ‘‘transver-

sus abdominis’’) OR MM ‘‘Abdominal Muscles1’’OR abdominal muscle*

9. (core) and (S2)10. (stabiliz*) and (S2)11. transverse abdominis OR TrA OR internal obliqu*

OR external obliqu*12. S8 or S9 or S10 or S1113. change* OR alter* OR improv*14. S7 AND S1315. S12 AND S1316. (((MM ‘‘Sciatica’’) OR ‘‘sciatica’’)) OR MM ‘‘Low

Back Pain’’ OR MM ‘‘Pelvic Pain1’’ OR MM‘‘Chronic Pain’’ OR MM ‘‘Back Pain1’’ OR lumbarpain OR back pain OR LBP OR dorsalgia ORlumbago OR lumbalgia OR backache

17. ‘‘course’’ OR (‘‘Predict*’’ OR (((MM ‘‘Risk Factors1’’)OR ‘‘risk factor’’)) OR (((MM ‘‘Morbidity1’’) OR‘‘morbidity’’ OR (MM ‘‘Comorbidity’’)) OR ((MM‘‘Prognosis1’’) OR ‘‘prognos*’’) OR ((MM ‘‘Inci-dence’’) OR ‘‘Incidence’’)))

18. ((MH ‘‘Randomized Controlled Trials’’) OR ‘‘-randomized controlled’’) OR ((((MM ‘‘ConcurrentProspective Studies’’) OR ‘‘cohort’’ OR (MM‘‘Nonconcurrent Prospective Studies’’)) OR longitu-dinal OR ((‘‘follow-up’’ OR (MM ‘‘ProspectiveStudies1’’)) OR followup)) OR longitudinal study

OR ((MM ‘‘Retrospective Design’’) OR ‘‘retrospec-tive’’) OR ‘‘cohort’’ OR cohort study)

19. S17 AND S1820. S14 AND S16 AND S1921. S15 AND S16 AND S19

SPORT Discus Search Strategies1. lumbar OR back OR trunk OR torso OR spine OR

lumbosacral OR Lumbar vertebrae OR coccy* ORiliac OR ilium

2. Musculoskeletal OR muscle*3. S1 and S24. MF5. (back extensor* OR multifid* OR LM) OR (paraspi-

nal AND (Musculoskeletal OR muscle*))6. S3 or S4 or S57. internal obliqu* OR external obliqu* OR TrA OR

(((rectus abdominis OR transversus abdominis ORtransverse abdominis) OR abdominal muscle* OR(abdomen AND (Musculoskeletal OR muscle*)))OR ((core AND (Musculoskeletal OR muscle*)) OR(stabiliz* AND (Musculoskeletal OR muscle*))))

8. change* OR alter* OR improv*9. S6 AND S810. S7 AND S811. backache OR sciatica OR LBP OR dorsalgia OR

lumbago OR lumbalgia12. pain AND (lumbar OR back OR trunk OR torso OR

spine OR lumbosacral OR Lumbar vertebrae ORcoccy* OR iliac OR ilium)

13. S11 OR S1214. retrospective OR prospective OR followup OR

follow-up OR follow up OR longitudinal15. ((predict* OR morbidity OR comorbidity OR prog-

nos* OR risk factor*) OR course) OR associat* ORcorrelat*

16. S14 AND S1517. S9 AND S13 AND S1618. S10 AND S13 AND S16Other search strategies for PEDro and Cochrane Library:(Low back pain OR LBP OR lumbago OR lumbalgia OR

backache) and (lumbar multifidus OR LM OR transversusabdominis OR TrAOR SM) and (prognos* OR predict* ORlongitudinal OR cohort OR course OR survival analysis ORfollow-up).

Wong et al The Journal of Pain 1.e27

Appendix II. Description of Studies That Investigated Temporal Changes in Morphology or Neuromuscular Function and Temporal Changes inClinical Outcomes Without Direct Calculation of Their Associations

STUDIES

STUDY POPULATION; AVERAGE

AGE AND STANDARD DEVIATION;FOLLOW-UP INTERVENTION TREATMENT PROCESS VARIABLES

OUTCOME

MEASURE COFOUNDER ADJUSTMENT RESULTS/EXPLAINED VARIANCE

Kankaanpaa

et al, 199957Patients with chronic LBP were

recruited from an

occupational health center

(n = 59); male in Group

A = 40.7 6 8.6 years; female

in Group B = 38.96 8.2 years;

male in Group

A = 38 6 6.9 years; female in

Group B = 40.6 6 8.1 years;

immediately after the

treatment program, 6 months

and 1 year

A randomized study.

A: 12-week physical exercises

twice weekly with a specific

trunk muscle–strengthening

device, stretching and

relaxation exercises,

behavioral support, and

ergonomic advice provided by

a physiotherapist

B: 4-week once-weekly

nonexercise treatment:

Massage therapy, thermal

therapy, medication

MPF slope of LM as measured

by surface EMG at L5-S1

level during an inertial

back extension endurance

test using a specific device

PDI

VAS

Analysis was adjusted for

the pretreatment

standard deviation of

a given dependent

variable.

No significant difference in PDI,

VAS, and MPF slope

(fatigability) between genders

in both treatment groups at

all time points.

There was statistically significant

improvement in PDI, VAS, and

MPF slope at all time points in

Group A.

No significant difference in PDI,

VAS, andMPF slope at all time

points in Group B.

Group A had significant

improvement of PDI and VAS

than Group B at all time

points.

Group A had significant

improvement of MPF slope

than Group B except at the 1-

year follow-up.

Sung, 200392 People with chronic LBP, without

detailed explanation of

recruitment process (n = 10

male and 6 female = 6);

47.9 6 12.2 years; 4 weeks

An uncontrolled study. Motor

control exercise for 3 times

per week for 4 weeks. The

exercise program included

upper-body extension,

alternate arm and leg lift in

prone, alternate arm and leg

extension in quadruped

position, curl-up with arms at

sides

Changes in median frequency

of LM during 1-minute

modified Sorenson test as

measured by surface EMG

ODI Analysis was adjusted

for age and onset time

of symptoms

ODI significantly decreased by

4.6 points (P-value was not

reported)

ANCOVA showed that LM

median frequency changed

from 136.4 6 30.2 Hz to

131.2 6 30.6 Hz (P < .05)

Significant interaction between

gender and treatment (F[1,

14] = 5.8; P = .031) with

significantly decreased LM

activity of men after

treatment (P < .045), whereas

women had nonsignificant

increase in LM activity

(P < .20)

1.e28

TheJournalofPain

Deep

TrunkMuscles

andBack

Pain—

ASystem

aticReview

Appendix II. Continued

STUDIES

STUDY POPULATION; AVERAGE

AGE AND STANDARD DEVIATION;FOLLOW-UP INTERVENTION TREATMENT PROCESS VARIABLES

OUTCOME

MEASURE COFOUNDER ADJUSTMENT RESULTS/EXPLAINED VARIANCE

Akbari et al, 20081 Participants with chronic LBP

were recruited from

physiotherapy clinics (n = 63);

Group A: 39.6 6 3.5 years;

Group B: 40 6 3.6 years;

8 weeks

A double-blind randomized

controlled trial with sessions

twice per week.

A: Motor control exercise for

LM and TrA cocontraction in

different positions

progressively.

B: General exercise

TrA/LM resting thickness as

measured by B-mode USI

VAS No adjustment for confounders Significant increase in

posttreatment TrA and LM

thickness (P < .0001) and

significant decrease in VAS in

both groups (P < .0001)

Group A had significant

decrease in VAS compared

with Group B (P = .015).

Improvement in the ratio

between posttreatment and

pretreatment TrA thickness in

Group A was greater than the

improvement ratio in the

Group B (P = .005).

Improvement in the ratio

between posttreatment and

pretreatment LM thickness in

Group A was greater than

improvement ratio in the

Group B (P = .004)

However, no significant

difference in improvement

ratio of VAS between the 2

groups.

Wonget

alTheJournalofPain

1.e29

Appendix II. Continued

STUDIES

STUDY POPULATION; AVERAGE

AGE AND STANDARD DEVIATION;FOLLOW-UP INTERVENTION TREATMENT PROCESS VARIABLES

OUTCOME

MEASURE COFOUNDER ADJUSTMENT RESULTS/EXPLAINED VARIANCE

Hides et al, 200842 Elite cricketers with LBP (n = 10);

21.2 6 2.0 years; 13 weeks

An uncontrolled trial. The 13-

week cricket-training

program included two 6-

week training blocks

separated by a 1-week break.

The first 6 weeks and the first

2 weeks of the second block

of the training camp involved

individual skill training, drills,

cardiovascular exercise,

weight training, and lectures.

The final 4 weeks involved

cricket matches.

Resting CSAs of bilateral LM

muscles from L2 to L5 by B-

mode USI

VAS Adjusted for age, height, and

body mass

There was 50.1% reduction of

VAS after 13-week program

There was no asymmetry in

bilateral LM from L2 to L4

levels at all time points

There was significant increase in

the CSAs of LM after

13 weeks.

At baseline, CSA of LM

ipsilateral to painful side at L5

level was 8.3% smaller than

the other side

After 13 weeks, no asymmetry

between bilateral sides of LM

at L5. CSA of LM on the side

ipsilateral to symptoms

increased by 26.2%, whereas

the contralateral side

increased by 20.7%

No correlation calculation

between changes in VAS and

changes in CSA

1.e30

TheJournalofPain

Deep

TrunkMuscles

andBack

Pain—

ASystem

aticReview

Appendix II. Continued

STUDIES

STUDY POPULATION; AVERAGE

AGE AND STANDARD DEVIATION;FOLLOW-UP INTERVENTION TREATMENT PROCESS VARIABLES

OUTCOME

MEASURE COFOUNDER ADJUSTMENT RESULTS/EXPLAINED VARIANCE

Marshall and

Murphy, 200876Participants with chronic

nonspecific LBP were

recruited by chiropractors and

physiotherapists (n = 50);

Group A: 34.3 6 9.2 years;

Group B: 35.8 6 10.4 years;

Group C: 33.9 6 9.6 years;

Group D: 41.7 6 10.7 years;

baseline, 4, 8, 16, and

56 weeks

A randomized controlled

trial. It involved a 16-week

intervention.

A: 4-week manipulative

treatment by chiropractors or

physiotherapists followed by

12-week progressive Swiss

ball exercise

B: 4-week manipulative

treatment by chiropractors or

physiotherapists followed by

home-based low back

exercises for 12 weeks

C: 4-week nonmanipulative

treatment by chiropractors or

physiotherapists followed by

12-week progressive Swiss

ball exercise

D: 4-week nonmanipulative

treatment by chiropractors or

physiotherapists followed by

home-based low back

exercises for 12 weeks

Changes in anticipatory

activation of bilateral TrA

during right rapid arm

movement as measured by

surface EMG

ODI No adjustment for confounders At the first 8 weeks, participants

in all groups had significant

reduction in ODI score than

baseline.

At 16 weeks, participants in

Group A and C had significant

reduction in ODI scores than

Group B and D (mean

difference = 7.7, P = .023)

From 16 weeks to 56 weeks,

patients in all groups had

significant decrease in latency

response of right TrA/IO

during rapid arm movements

No significant change in the

latency responses measured

during the abduction and

extension directions.

There was no significant change

in the latency of left TrA/IO

onset.

Franca et al,

2010; 201223,27Patients with chronic

nonspecific LBP without leg

pain were recruited from the

orthopedic department of a

hospital (n = 30); Group A:

42.076 8.15 years; Group B:

41.736 6.42 years; Group C:

41.53 6 4.41 years; 6 weeks

A randomized controlled trial.

Each group received 12

treatment sessions over

6 weeks (30 minutes each):

A: Motor control group:

training function of TrA and

LM in lying positions, and

contraction of TrA and LM in

upright position.

B: Superficial strengthening

group: strengthening

exercises for RA, OI, OE, ES

C: Stretching group:

stretching of erector spinae,

hamstrings, triceps sure in

different positions

Change in TrA activation

capacity as measured by the

change pressure of PBU in

prone patients

VAS

MPQ

ODI

No adjustment for any

confounders

Group A: VAS, MPQ, and ODI

scores improved by 90–99%,

whereas TrA contraction as

measured by PBU improved

by 48.32%.

Group B: VAS, pain-McGill and

ODI scores improved by 43–

61%, whereas TrA

contraction as measured by

PBU decreased by 5.11%.

Group C: VAS, pain-McGill and

ODI scores improved by 37–

52%, whereas TrA

contraction as measured by

PBU improved by 6.56%.

Wonget

alTheJournalofPain

1.e31

Appendix II. Continued

STUDIES

STUDY POPULATION; AVERAGE

AGE AND STANDARD DEVIATION;FOLLOW-UP INTERVENTION TREATMENT PROCESS VARIABLES

OUTCOME

MEASURE COFOUNDER ADJUSTMENT RESULTS/EXPLAINED VARIANCE

Kumar et al, 201066 Patients with subacute/chronic

LBP from a physician in the

department of physical

medicine and rehabilitation at

a university (n = 141); Group

A: 35.8 6 .7 years; Group B:

34.4 6 .7 years; Day 10, Day

20, 3 months and 6 months

A randomized controlled trial.

Each group received 20

treatments over 20 days

A: Conventional

physiotherapy group:

ultrasound, short wave

diathermy, lumbar

strengthening exercises,

sitting

B: Dynamic muscular

stabilization group:

progressive ADIM, abdominal

hollowing in different

positions and limb

movements

Back pressure changes and

abdominal pressure changes

using PBU

VAS

SF-36

No adjustment for any

confounders

All ANOVAs for all outcome

variables showed no

significant difference

between sexes.

Themean VAS and SF-36 scores,

back pressure changes, and

abdominal pressure changes

showed that Group A had

significantly better

improvement than Group B.

Coghlan et al, 201113 Participants with chronic LBP

between L3 and the gluteal

fold without leg pain were

recruited by an advertisement

in a newspaper (n = 13);

38.9 6 9.2 years; 6 weeks

An uncontrolled study.

Participants received 15 to

30 minutes neuromuscular

electrical stimulation of

lumbar paraspinal muscles

and anterolateral abdominal

muscles at a frequency of 1 to

2 times a day for 6 weeks

Change in bilateral TrA and OI

percent thickness during

ASLR; Change in bilateral LM

percent thickness change

during CALT as measured by

B-mode USI

VAS No adjustment for

confounders

There were significant increases

in left and right TrA and OI

thickness during ASLR

(P < .05).

VAS scores over last 2 weeks

were significantly improved at

the end of treatment

(P < .001)

Lee et al, 201168 Patients with chronic LBP from a

hospital (n = 33); Group A:

32.7 6 5.9 years; Group B:

33.1 6 5.7 years; 12 weeks

Participants received either type

of 12-week treatment at 3

times/week

A: Gymnastik ball exercise

therapy: stretching, ball push

up, bridging, crunch, ball sit

up exercise, etc (45 minutes

per visit)

B: Conventional physical

therapy treatment:

ultrasound therapy, heat

treatment, and

transcutaneous electrical

nerve stimulation (50 minutes

per visit).

CSA of LM at L4/5 using

CT scan

VAS No adjustment for

confounders

VAS scores of both groups were

statistically significantly

improved after treatments.

Mean CSA of LM was

significantly increased in

Group A but there was no

significant change in CSA of

LM in Group B.

Mean CSA of LM in Group A

was significantly larger than

that in Group B following

treatments.

1.e32

TheJournalofPain

Deep

TrunkMuscles

andBack

Pain—

ASystem

aticReview

Appendix II. Continued

STUDIES

STUDY POPULATION; AVERAGE

AGE AND STANDARD DEVIATION;FOLLOW-UP INTERVENTION TREATMENT PROCESS VARIABLES

OUTCOME

MEASURE COFOUNDER ADJUSTMENT RESULTS/EXPLAINED VARIANCE

Williams et al, 2013112 Volunteers with acute (n = 20)

and chronic LBP (n = 20) were

recruited from physiotherapy

and chiropractic clinics;

42.7 6 6.8 years for acute

LBP; 36.6 6 10.8 years for

chronic LBP; within a session

An uncontrolled study.

Participants took different

brands of oral analgesia.

Onset time, peak activity and

EMG profile of LM as

measured by surface EMG

during trunk flexion,

extension, side flexion and

rotation of lumbar, and a box

lift

Worst pain

VAS during

each movement

or task

No adjustment for confounders VAS scores were significantly

reduced in flexion, extension,

and lifting.

The onset time of LM (except for

the left LM of volunteers with

LBP during the lifting task)

during trunk flexion, left-side

flexion, right-side flexion, and

lifting was not significantly

changed following immediate

pain reduction.

The peak EMG activities of LM

during flexion, left-side

flexion, and the lifting tasks

were not significantly

changed.

EMG profile of LM in different

tasks changed variably

following partial pain relief

Wonget

alTheJournalofPain

1.e33

Appendix II. Continued

STUDIES

STUDY POPULATION; AVERAGE

AGE AND STANDARD DEVIATION;FOLLOW-UP INTERVENTION TREATMENT PROCESS VARIABLES

OUTCOME

MEASURE COFOUNDER ADJUSTMENT RESULTS/EXPLAINED VARIANCE

Zielinski et al, 2013116 Participants with chronic LBP

with or without recurrences

recruited from local physical

therapy and physician clinics

and/or from self-referral

(n = 25); 34.36 10.8 years for

CPR-eligible patients and

42.1 6 10 years for CPR-

ineligible patients; 6 weeks

One arm of a randomized

controlled study. Each

participants received 1

treatment session per week

over 6 weeks

Motor control exercise:

training function of specific

deep muscles of lumbar

region, strengthening trunk

flexors, extensors, and

oblique muscles, education

on proper body biomechanics

and spinal protection during

activities of daily living, home

exercises

Percent thickness change of

LM thickness during a

resisted or nonresisted

CALT in prone as measured

by B-mode USI

Modified

ODI;

NPRS

No adjustment for

cofounders

LM thickness change following

treatment:

In CPR-eligible patients

No resistance task: No

significant change (P = .08)

Resistance task: No

significant change (P = .26)

In CPR-ineligible patients

No resistance task: increased

from 10.8 to 14.7%

(P = .05)

Resistance task: No

significant change (P = .26)

Reduction in ODI following

treatment

CPR-eligible group: from

25.3 6 7.7% to

14.6 6 8.8% (P = .003)

CPR-ineligible group: from

14.6 6 5.5% to

10.3 6 6.6% (P = .10)

Reduction in NPRS following

treatment

CPR-eligible group: from

2.8 6 1.7 to 2.0 6 1.5

(P = .13)

CPR-ineligible group: from

1.8 6 1.3 to 1.4 6 1.3

(P = .17)

Abbreviations: MPF, mean power frequency; EMG, electromyography; PDI, pain and disability index; VAS, visual analog scale; ODI, Oswestry Disability Index; ANCOVA, analysis of covariance; B-mode, brightness-mode (ultrasound);

CSA, cross-sectional area; RA, rectus abdominis; OI, obliquus internus; OE, obliquus externus; ES, erector spinae; PBU, pressure biofeedback unit; MPQ, McGill Pain questionnaire; ADIM, abdominal drawing-in maneuver; SF-36, Short

Form 36 quality of life questionnaire; ANOVA, analysis of variance; ASLR, active straight leg raise test; CALT, contralateral arm lifting task; CT, computed tomography; CPR, clinical prediction rule; CPR-eligible patients, patients who either

had lumbar spine hypomobility and/or had any 3 of 4 clinical features identified by the clinical prediction rule for stabilization exercise success; CPR-ineligible patients, patients who neither had lumbar spine hypomobility nor had any 3 of 4

clinical features identified by the clinical prediction rule for stabilization exercise success; NPRS, numeral pain rating scale.

1.e34

TheJournalofPain

Deep

TrunkMuscles

andBack

Pain—

ASystem

aticReview

Appendix III. Inclusion and Exclusion Criteria

INCLUSION CRITERIA

Type of study design: Systematic review/meta-analysis of cohort studies or follow-up studies; validation of clinical decision rule; follow-up of

untreated control groups in randomized controlled trials; retrospective cohort studies; observational studies; case series; nonduplicated

information from multiple reports of the same patient cohorts

Articles: English, French, Chinese, or Portuguese language; non-English, non-Chinese, and non-Portuguese language studies in English-language

review articles; inclusion of morphology measurement or muscle activity of LM or TrA, and at least 1 clinical outcome of LBP.

Subjects: Self-ambulatory adult patients (>18 years of age) with acute (<6 weeks), subacute (6 to 12 weeks), or chronic (>12 weeks) nonspecific

LBP115

Personal characteristics: human subjects of any gender, BMI, and ethnicity

Causes of pain: idiopathic; traumatic without vertebral fracture, degenerative; experimentally induced

Setting: in-patient; outpatient clinics; laboratories

Treatment: nonsurgical treatment with explicit description of therapeutic dosage (eg, drugs, physical therapy modalities, or manual therapy), or

studies without intervention

Follow-up: at least 1 follow-up after the baseline measurements (with reported timing between follow-up)

Prognostic or treatment effect modifiers:

Morphology or neuromuscular function of LM or TrA

Type of outcome measures:

At least 1 of the following LBP clinical outcome measures: Quebec Back Pain Disability scale; Roland Morris disability

questionnaire (RMDQ); Oswestry Disability Index (ODI); Japanese orthopedic association score; McGill pain questionnaire; Patient-Specific

Functional Scale; pain score (eg, visual analog scale, numerical pain rating scale); subjective improvement of symptoms (global rating of change);

return-to-work/school; days off work/school; recurrent rate; quality of life (eg, short-form 36, short-form 12, EQ-5D, etc); subjective complaints

LM or TrA measures: magnetic resonance imaging; computed tomography; ultrasound imaging; electromyography (intramuscular or

surface)

Measurement period: multiple assessments over different time periods in prospective or retrospective studies

Statistical analysis: odds ratio; risk ratio; linear or logistic regression; multiple regression; cross tabulations; chi-squared test; Kaplan-Meier

curves; Cox regression; ANOVA; Wilcoxon rank sum test; Kruskal-Wallis test

EXCLUSION CRITERIA

Designs: single cross-sectional studies;multiple reports that included duplicated results of same patient case series (only include the studywith the

largest series)

Medical conditions: pregnancy related back pain; LBP patients from sources outside the back (nonspinal LBP); fibromyalgia or myofascial pain

syndrome in >20% of the overall sample size; spondylolisthesis; spondylolysis; stenosis; disc herniation; back surgery; cancer; neural disease (eg,

multiple sclerosis or Guillain-Barr�e syndrome); acute major trauma; spinal cord injury; congenital lumbar/sacral condition; spasticity resulting

from spinal cord injury or neural diseases; cauda equina syndrome; infection; solely osteoporosis; vertebral compression fracture; rheumatoid

arthritis; scoliosis; stroke; cerebral palsy

Articles: papers written in languages other than English, Chinese, French, and Portuguese; letters; studies in books; commentaries

Wong et al The Journal of Pain 1.e35