Manual Therapy 14 (2009) 496–500

Contents lists available at ScienceDirect

Manual Therapy
journal homepage: www.elsevier.com/math

Original Article

The effect of chronic low back pain on size and contraction

of the lumbar multifidus muscleq
Tracy L. Wallwork a, b, Warren R. Stanton c, Matt Freke d, Julie A. Hides a, c, *
a
Division of Physiotherapy, School of Health and Rehabilitation Sciences, The University of Queensland, Brisbane, 4072, Australia
b
Private practice, Perth, Australia
c
UQ/Mater Back Stability Clinic, Mater Health Services, South Brisbane, Qld, 4101, Australia
d
Physiotherapy department, Second Health Support Battalion, Gallipoli Barracks, Enoggera, Brisbane, Queensland, Australia

a r t i c l e i n f o a b s t r a c t

Article history: Decreases in the size of the multifidus muscle have been consistently documented in people with low
Received 22 April 2008 back pain. Recently, ultrasound imaging techniques have been used to measure contraction size of the
Received in revised form multifidus muscle, via comparison of the thickness of the muscle at rest and on contraction. The aim of
26 August 2008
this study was to compare both the size (cross-sectional area, CSA) and the ability to voluntarily perform
Accepted 25 September 2008
an isometric contraction of the multifidus muscle at four vertebral levels in 34 subjects with and without
chronic low back pain (CLBP). Ultrasound imaging was used for assessments, conducted by independent
keywords:
examiners. Results showed a significantly smaller CSA of the multifidus muscle for the subjects in the
Chronic low back pain (CLBP)
Multifidus muscle CLBP group compared with subjects from the healthy group at the L5 vertebral level (F ¼ 29.1, p ¼ 0.001)
Ultrasound imaging and a significantly smaller percent thickness contraction for subjects of the CLBP group at the same
vertebral level (F ¼ 6.6, p ¼ 0.02). This result was not present at other vertebral levels (p > 0.05). The
results of this study support previous findings that the pattern of multifidus muscle atrophy in CLBP
patients is localized rather than generalized but also provided evidence of a corresponding reduced
ability to voluntarily contract the atrophied muscle.
Ó 2008 Elsevier Ltd. All rights reserved.

1. Introduction and increased incrementally in size on progression caudally (Hides

et al., 1995). The cross-sectional area (CSA) of the multifidus muscle
The lumbar multifidus muscle has been the subject of consid- has been shown to be larger in male subjects, and age was not
erable research using imaging techniques including magnetic related to multifidus size (Stokes et al., 2005).
resonance imaging (MRI), computerised tomography (CT scanning) The size of the multifidus muscle has also been documented in
and ultrasound imaging (for review, see Stokes et al., 2007). Two people with acute and chronic LBP. In patients with acute unilateral
aspects of muscle function that can be assessed using imaging LBP, studies employing ultrasound imaging demonstrated
techniques are muscle size (MRI, CT, ultrasound imaging) and a decrease in the CSA of the multifidus muscle ipsilateral to pain-
muscle contraction (ultrasound imaging). The clinical relevance of ful symptoms (Hides et al., 1994, 1996), and the atrophy was
these techniques is that they allow documentation of morphology predominantly isolated to one vertebral level (L5). Multifidus
and dynamic muscle function in both healthy subjects and those muscle atrophy has been quantified in MRI and CT studies in terms
with acute and chronic low back pain (CLBP). Detection of changes of both decreased muscle size and alterations in muscle consistency
in multifidus muscle size and motor control in people with low (due to fatty deposits or fibrous/connective tissue infiltration). In
back pain (LBP) (by comparison with healthy subjects) may provide a CT study, Danneels et al. (2000) showed that the multifidus
useful information which can be used to guide rehabilitation muscle was selectively decreased at the lowest lumbar level in
approaches. patients with CLBP when compared with control subjects. In an
In healthy subjects, the lumbar multifidus muscle has been MRI study of 78 patients with CLBP, degeneration of the multifidus
shown to be symmetrical between sides (Hides et al., 1994, 1995) muscle was present in 80% of the participants, and most commonly
was seen at L4–5 and L5–S1 (Kader et al., 2000). Similar patterns of
q Work attributed to: Division of Physiotherapy, School of Health and Rehabili-
atrophy have been demonstrated using ultrasound imaging (Hides
tation Sciences, The University of Queensland, Brisbane, 4072, Australia.
et al., 2008a,b).
* Corresponding author. Tel.: þ61 7 3365 2718; fax: þ61 7 3365 1877. Dynamic studies of the multifidus muscle are of interest as they
E-mail address: [email protected] (J.A. Hides). provide information related to motor control of the muscle.

1356-689X/$ – see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.math.2008.09.006
 T.L. Wallwork et al. / Manual Therapy 14 (2009) 496–500 497

Viewing the muscle contracting in parasagittal section, two studies recorded for all subjects. Examiner 2 assessed multifidus muscle
have provided feedback of multifidus muscle contraction to size (CSA) and examiner 3 assessed the ability to voluntarily
patients with acute (Hides et al., 1996) and chronic LBP (Hides et al., contract the multifidus muscle (muscle thickness measures).
2008a). A randomized controlled trial conducted by Van et al. Examiners 2 and 3 were blinded to group allocation of the subjects.
(2006) showed that provision of visual biofeedback using ultra- The assessors were blinded to each other’s results.
sound imaging improved the ability to contract the multifidus
muscle in healthy subjects. 2.2.1. Assessment using ultrasound imaging
While observation of multifidus muscle contraction using Ultrasound imaging was conducted using Diasonics Synergy
ultrasound imaging has been used in clinical practice for quite ultrasound imaging apparatus equipped with a 5 MHz curvilinear
some time, it is only recently that the technique has been validated transducer (GE-Diasonics, Japan). Subjects were positioned in prone
by comparison with fine wire electromyography (EMG). Kiesel et al. lying, with a pillow placed under the abdomen to minimize the
(2007) used graded resistance of contralateral upper extremity lifts lumbar lordosis. The spinous processes from L2–L5 were marked
to produce incremental involuntary contraction of the lumbar with a pen. Detection of spinous processes was determined manu-
multifidus muscles and demonstrated a relationship between ally using the iliac crests as a landmark. The location of the spinous
increases in muscle thickness and fine wire EMG activity for processes was then confirmed using ultrasound imaging by viewing
contractions of 19–43% of maximum. In addition, the effect of pain the spinous processes relative to the sacrum in sagittal section.
on multifidus muscle function has been demonstrated experi-
mentally by using a model of induced pain (Kiesel et al., 2008). 2.2.2. Assessment of multifidus muscle CSAs
However, the ability of patients with CLBP to voluntarily contract CSAs of the multifidus muscle were measured from L2 to L5
the multifidus has not been formally assessed. vertebral segments. Reliability of performing these measures has
The aim of this study was to compare both the size (CSAs) and been previously reported (Hides et al., 1992, 1994; Stokes et al., 2005;
the ability to voluntarily contract the multifidus muscle at four Pressler et al., 2006) and previous clinical trials have shown the
vertebral levels in subjects with and without CLBP using real-time highly trained assessor (examiner 2) in the present study to be
ultrasound imaging. repeatable and reliable with ultrasound measurements of multifidus
muscle CSA (Hides et al., 1992, 1994). The validity of measurements
2. Methods obtained using ultrasound imaging has also been demonstrated by
comparison with MRI measurements (Hides et al., 1995).
2.1. Subjects Subjects were instructed to relax the paraspinal musculature,
electroconductive gel was applied, and the transducer placed
Seventeen subjects with CLBP (8 males, 9 females, age range transversely over the spinous process of the vertebral level being
18–60) and seventeen healthy subjects (8 males, 9 females, age measured. This produced images in which the spinous process and
range 18–45) participated in the study (Table 1). CLBP was defined laminae could be seen, with multifidus muscles visible on both
in this study as a history of non-specific LBP for more than 3 months sides of the spine (Fig. 1A). The echogenic vertebral lamina was
(International Association for the Study of Pain, 2008). Exclusion used consistently as a landmark to identify the muscle’s deep
criteria for the CLBP subjects included histories suggestive of non- border. The multifidus muscle is bordered superiorly by the thor-
mechanical LBP, overt neurological signs, previous lumbar surgery, acolumbar fascia, and the medial border was provided by the
self-reported pain levels of less than 3 on a visual analogue scale acoustic shadow from the tip of the spinous process of the vertebral
(VAS) and LBP associated with a worker’s compensation or motor level being assessed. The lateral border was formed by the fascia
vehicle accident claim. Exclusion criteria for all subjects included surrounding the multifidus and separating it from the longissimus
pregnancy, presence of spinal abnormalities, presence of scoliosis component of the lumbar erector spinae muscle.
with a rib height difference of greater than 2 cm on forward flexion, Bilateral images of the multifidus muscles were obtained where
histories of severe trauma, spinal or abdominal surgery, reported possible (Fig. 1A), except in the case of larger muscles where left
neuromuscular or joint disease, training involving the back muscles and right sides were imaged separately. The CSA (in cm2) of the
within 3 months and difficulty lying in the prone position. This multifidus was measured by tracing around the muscle border with
study was approved by the Medical Research Ethics Committee at the on-screen cursor (Fig. 1B). For consistency, the inner edge of the
the host institution. Informed consent was obtained and the rights border was used.
of human subjects were protected.
2.2.3. Assessment of multifidus muscle thickness and contraction
Prior to testing of contraction of the multifidus muscle, all
2.2. Procedure
subjects received an initial explanation. The anatomical location of
the multifidus muscle was demonstrated using a model of the
Three experienced musculoskeletal physiotherapists were
lumbar spine, and pictures of the muscle were provided and
involved in data collection. Examiner 1 was responsible for
explained. A demonstration of an isometric contraction of the
applying inclusion and exclusion criteria and collection of demo-
biceps was performed as a simple example of the type of contrac-
graphic data. For subjects with LBP a body chart was used to record
tion required. Subjects were further instructed to take a relaxed
distribution of symptoms and a VAS was used to assess pain levels
breath in and out, pause breathing and then try to ‘‘swell’’ or
experienced over the last week. Height and weight were measured
contract the muscle. They were also instructed not to move their
and age, gender and weekly activity levels (<1.5, 1.5–3, >3 h) were
spine or pelvis when they contracted the muscle, and the type of
muscle contraction required was a slow gentle sustained contrac-
Table 1 tion. To familiarize subjects with the contraction prior to measuring,
Demographics of subjects in Group 1 (CLBP) and Group 2 (Unimpaired). subjects were asked to perform 3 contractions with tactile and
verbal feedback while the examiner manually palpated the multi-
Age Weight Height
fidus muscle. It was explained to the subjects that during testing
Mean (SD) Mean (SD) Mean (SD)
they would have 5 s to try to contract the multifidus muscle and
Group 1 (Unimpaired) n ¼ 17 33.9 (11.2) 81.2 (12.5) 176.6 (10.3) hold the contraction. At the end of the 5 s period, the image would
Group 2 (CLBP) n ¼ 17 41.9 (13.7) 76.1 (16.7) 174.2 (10.3)
be saved on the ultrasound screen, and measurements performed.
498 T.L. Wallwork et al. / Manual Therapy 14 (2009) 496–500

Fig. 1. A. Bilateral transverse ultrasound image at the L4 vertebral level, without CSA tracings. B. Bilateral transverse ultrasound image at the L4 vertebral level with CSA tracings.
The CSA (in cm2) of the multifidus was measured by tracing around the muscle border with the on-screen cursor.

The multifidus muscle was imaged in parasagittal (longitudinal) address this issue, the data for two males and two females were
section allowing visualisation of the zygapophyseal joints, muscle excluded from the study as outliers (more than 3 standard devia-
bulk and thoracolumbar fascia (Hides et al., 1992, 1995, 1996; Van tions above the sample mean). In the analyses of the thickness
et al., 2006). The multifidus muscle was imaged on both sides from contraction, the data for 30 participants were used (16 in the CLBP
L2 to L5 vertebral levels. Linear measurements (multifidus muscle group and 14 in the unimpaired group). Due to limited availability
thickness measures) using on-screen callipers were made in all of examiner 2, CSA of the multifidus muscle was only possible for
cases from the tip of the zygapophyseal joint to the superior border 22 of the 30 participants (11 in the CLBP group and 11 in the
of the multifidus muscle for each vertebral level (Fig. 2). unimpaired group).
In order to assess multifidus muscle contraction, the difference A mixed design analysis of covariance (ANCOVA) was used
between the multifidus muscle thickness at rest and during to separately analyse the outcome measures of ‘percent change in
contraction was calculated. A split-screen technique was used to multifidus muscle thickness due to contraction’ (called multifidus CSA
make this measurement more reliable, by allowing anatomical percent thickness contraction) at each vertebral level. Percent thick-
orientation to be maintained in both cases (Fig. 2). Subjects were ness contraction was calculated as; [(contracted thickness  resting
not allowed to watch the ultrasound monitor or receive feedback thickness/contracted thickness)  100]. In this study there were 7
about the contractions performed during testing. independent variables: age, weight, height, gender, activity level
Prior to the present study, a reliability trial was performed on 10 (coded as low, moderate or high), group (CLBP or unimpaired), and the
healthy subjects not involved in the main study (Wallwork et al., repeated measures of asymmetry (larger or smaller side). The variables
2007). Each subject was positioned in the standard testing position. of age, weight and height were treated as covariates in the analyses.
Three separate ultrasound images were obtained at rest and the Post-hoc contrasts were used to test for differences among the 3
anteroposterior (thickness) measurement was conducted on para- activity levels if the main effect for this factor was statistically
sagittal images at two vertebral levels by two raters (examiner 3 significant.
and an expert). Intraclass correlations (ICC) were used to determine For both the dependent variables, measures of ‘size’ and
intra-rater and inter-rater reliability. Results of the ICC3,1 for intra- ‘asymmetry’ are of interest. As calculation of average size across
rater reliability was 0.89 for L2/3 (95%CI ¼ 0.72–0.97) and 0.88 for ‘side’ is confounded by asymmetry across ‘side’, the data for ‘larger’
L4/5 (95%CI ¼ 0.68–0.97). The results of the ICC2,3 for inter-rater and ‘smaller’ side were used rather than calculating the percentage
reliability was 0.96 for L2/3 (95%CI ¼ 0.84–0.99) and 0.97 for L4/5 difference between the ‘larger’ and ‘smaller’ sides. In addition, as
(95%CI ¼ 0.87–0.99) (Wallwork et al., 2007). higher-order interactions between the covariates and factors
confounded the analysis, a Type I sums of squares model was used
in preference to a Type III model.
2.3. Statistical analysis

Those without low back pain have greater capacity to produce 3. Results
relatively larger contractions than those with low back pain,
threatening the homogeneity of variance of the two samples. To Demographic details for subjects of both groups are shown in
Table 1. Results of an initial analyses of variance showed that there
was no significant difference between the two groups for the
variables of age (F ¼ 3.4, p ¼ 0.07), weight (F ¼ 1.0, p ¼ 0.32) and
height (F ¼ 0.5, p ¼ 0.49). Results of a chi-square test showed that
both genders (p ¼ 0.63) and all 3 activity levels (p ¼ 0.75) were
represented in similar proportions across the 2 groups.

3.1. Multifidus muscle size

Results of the ANCOVA showed a significantly smaller CSA of the

multifidus muscle for the CLBP group compared to the unimpaired
group at the L5 vertebral level (F ¼ 29.1, p ¼ 0.001), and slightly
larger size at L2 (F ¼ 5.8, p ¼ 0.047). A small but significant effect
Fig. 2. Ultrasound image of the multifidus muscle in parasagittal section at rest and on
(mean net difference of 0.17 cm2) was found for multifidus muscle
contraction using a split-screen. Relaxed multifidus muscle thickness ¼ 2.68 cm, con- ‘asymmetry’ at each vertebral level (p < 0.05) but this was similar
tracted value ¼ 3.20 cm. for both groups (p > 0.05). There were no significant effects for
 T.L. Wallwork et al. / Manual Therapy 14 (2009) 496–500 499

‘gender’ at the L4 and L5 vertebral levels, but male subjects had Table 3
significantly larger mutifidus muscles than females at the L3 Activity level differences (marginal meansa and standard deviations) in multifidus
size (cm2) across vertebral levels L2–L5.
vertebral level (difference of 0.9 cm2) and at L2 (difference of
0.15 cm2). ‘Activity level’ was significantly associated with multi- Physical activity per week L2 L3 L4b L5b
fidus size at L3 (F ¼ 5.9, p ¼ 0.03), L4 (F ¼ 11.9, p ¼ 0.006) and L5 Mean (SD) Mean (SD) Mean (SD) Mean (SD)
(F ¼ 5.4, p ¼ 0.04), but this was similar for both groups (no inter- Level 1 activity (<1.5 h) 1.78 (1.1) 2.24 (1.5) 2.95 (1.2) 3.96 (1.3)
action effect, p > 0.05). There were no significant higher-order Level 2 activity (1.5–3 h) 2.27 (0.9) 3.09 (1.3) 4.29 (1.1) 5.26 (1.2)
interactions. Table 2 shows the estimated marginal means of Level 3 activity (>3 h) 2.44 (1.3) 3.82 (1.9) 4.87 (1.5) 4.84 (1.7)
multifidus CSA for the CLBP and unimpaired groups at each level. a
Adjusted for the covariates of age, weight and height.
b
Table 3 shows the estimated marginal means of multifidus size for Statistically significant difference between Activity Level 1 versus Level 2 and
the 3 levels of activity. Level 3, based on post-hoc contrasts with Bonferroni correction.

3.2. Multifidus muscle thickness and contraction that rehabilitation may need to be specific in order to target
localized impairments in motor control.
Table 4 shows the thickness measurements of the multifidus Clinical approaches targeting motor control of muscles
muscle for rest and contracted conditions, averaged across left and including the multifidus, transversus abdominis and pelvic floor
right sides. Analysis of these data was based on calculation of the have been shown to be effective in randomized clinical trials (RCTs)
percent contraction from rest. Results of the ANCOVA showed (Hides et al., 1996; O’Sullivan et al., 1997; Stuge et al., 2004; Goldby
a significantly smaller percent thickness contraction for the CLBP et al., 2006). A RCT conducted on subjects with first episode acute
group compared to the unimpaired group at the L5 vertebral level LBP provided the first evidence of a localized, segmental impair-
(F ¼ 6.6, p ¼ 0.02), but not at other vertebral levels (p > 0.05). ment in the CSA of the multifidus muscle (Hides et al., 1996). Similar
A small but significant effect (mean net difference of 2.7%) was to the findings of the current study, it was reported that subjects
found for contraction ‘asymmetry’ at each vertebral level (p < 0.05) could not voluntarily contract the multifidus muscle at the verte-
but this was similar for both groups (i.e. there was no significant bral level where the atrophy of the muscle was observed. A tailored
interaction between ‘asymmetry’ and ‘group’, p > 0.05). There were exercise approach targeting the impaired muscle restored muscle
no significant effects for the variables of ‘activity level’ or ‘gender’ size and resulted in lower recurrence rates of LBP (Hides et al.,
and no significant higher-order interactions. Table 5 shows the 2001). Ultrasound imaging was used to provide feedback of mul-
estimated marginal means (and standard deviations) for the CLBP tifidus muscle contraction (Hides et al., 1996; Van et al., 2006).
and unimpaired groups at each level. A motor control approach was also recently successfully employed
in a study involving elite cricketers with LBP (Hides et al., 2008b).
4. Discussion Results showed that the CSAs of the multifidus muscles at the L5
vertebral level increased with training and these changes were
4.1. Multifidus muscle size commensurate with a 50% decrease in mean reported pain levels.
The finding that subjects who have LBP are less able to contract
Results from the current investigation showed a specific and the multifidus has also been reported in a laboratory study. The effect
localized pattern of atrophy of the multifidus muscles in the pres- of pain on multifidus muscle function was demonstrated experi-
ence of chronic LBP. In this study, atrophy was greatest at the L5 mentally using a model of induced pain (Kiesel et al., 2008). Increases
vertebral level, and there was a trend towards significance at the L4 in multifidus muscle thickness during arm lifting tasks were signif-
vertebral level. Several previous imaging studies have reported icantly reduced by pain in response to injection of saline into the
evidence of multifidus muscle atrophy in patients with LBP. erector spinae muscles. While Kiesel et al. (2008) did not examine
Researchers have investigated post-operative patients (Sihvonen voluntary contractions of the multifidus muscle, the findings may
et al., 1993), patients with acute/subacute LBP (Hides et al., 1994, support the current clinical practice of using physiotherapeutic
1996) and patients with chronic LBP (Kader et al., 2000; Danneels modalities to decrease pain prior to commencing rehabilitation of the
et al., 2000, 2001; Barker et al., 2004; Hides et al., 2008a,b). In multifidus muscle, and performance of voluntary multifidus
agreement with these previous studies, the pattern of atrophy seen contractions in pain-free positions (Hides et al., 1996).
in the chronic LBP patients investigated appeared to be specific and
localized in nature. 4.2.1. Limitations and future directions
This study has some limitations. The study sample size is small,
4.2. Multifidus muscle thickness and contraction though comparable with other similar investigations (Hides et al.,
1996; Danneels et al., 2000; Van et al., 2006). Thickness measures
The results of this study suggest that the neuromotor control of of the multifidus muscle were obtained in 30 subjects, where CSA
multifidus was altered at the L5 vertebral in patients with CLBP. measures were only obtained in 22 participants. While this is not
Subjects with CLBP were less able than healthy subjects to volun- ideal, the results from this study in relation to CSA of the multifidus
tarily contract the multifidus muscle at the same vertebral level are in line with previous reports (e.g. Danneels et al., 2000; Hides
where atrophy was present. The clinical relevance of this finding is
Table 4
Thickness (means and standard deviations) of the multifidus muscle in the rest and
Table 2 contracted state across vertebral levels L2–L5 (mm).
Group differences (marginal meansa and standard deviations) in multifidus muscle
L2 L3 L4 L5
size (cm2) across vertebral levels L2–L5.
Mean (SD) Mean (SD) Mean (SD) Mean (SD)
L2b L3 L4 L5b
Group 1 (unimpaired)
Mean (SD) Mean (SD) Mean (SD) Mean (SD) Rest 29 (5.2) 33 (5.0) 35.9 (5.3) 35.9 (4.8)
Group 1 (unimpaired) 1.94 (0.9) 3.09 (1.3) 4.61 (1.0) 5.56 (1.1) Contracted 31 (5.3) 34.7 (4.8) 37.7 (4.9) 38.1 (4.8)
Group 2 (CLBP) 2.40 (0.9) 3.02 (1.4) 3.47 (1.1) 3.81 (1.2) Group 2 (CLBP)
a
Rest 27.6 (4.7) 30.5 (4.5) 33.6 (5.3) 33.9 (5.5)
Adjusted for the covariates of age, weight and height. Contracted 28.9 (4.7) 31.9 (5.2) 34.6 (5.4) 35.0 (5.6)
b
Statistically significant difference at p < 0.05.
500 T.L. Wallwork et al. / Manual Therapy 14 (2009) 496–500

Table 5 Danneels L, Van der Straeten G, Cambier D, Witvrouw E, De Cuyper H. The effects of
Group differences (marginal meansa and standard deviations) in percentage thick- three different training modalities on the cross-sectional area of the lumbar
ness contraction across vertebral levels L2–L5. multifidus. British Journal of Sports Medicine 2001;35:186–94.
Goldby LJ, Moore AP, Doust J, Trew M. A randomized controlled trial investigating
L2 L3 L4 L5b the efficiency of musculoskeletal physiotherapy on chronic low back disorder.
Spine 2006;31(10):1083–93.
Mean (SD) Mean (SD) Mean (SD) Mean (SD) Hides JA, Cooper DH, Stokes MJ. Diagnostic ultrasound imaging for measurement of
Group 1 (unimpaired) 6.93 (6.7) 5.20 (7.0) 5.15 (5.9) 6.29 (6.5) the lumbar multifidus in normal young adults. Physiotherapy Theory and
Group 2 (CLBP) 4.10 (7.4) 4.22 (7.8) 2.93 (6.5) 3.05 (7.2) Practice 1992;8:19–26.
Hides J, Gilmore C, Stanton W, Bohlscheid E. Multifidus size and symmetry among
a
Adjusted for the covariates of age, weight and height. chronic LBP and healthy asymptomatic subject. Manual Therapy 2008a;13(1):
b
Statistically significant difference at p < 0.05. 43–9.
Hides J, Jull G, Richardson C. Long-term effects of specific stabilizing exercises for
first-episode low back pain. Spine 2001;26(11):E243–8.
et al., 2008a). Furthermore, while CSA of the multifidus muscle was Hides J, Richardson C, Jull G. Multifidus recovery is not automatic after resolution of
measured, consistency changes in the muscle (fatty deposits or acute, first-episode low back pain. Spine 1996;21:2763–9.
fibrous/connective tissue infiltration) were not assessed. Future Hides J, Richardson C, Jull G. Magnetic resonance imaging and ultrasonongraphy of
the lumbar multifidus muscle: comparison of two different modalities. Spine
studies, especially those employing imaging techniques such as CT 1995;20(1):54–8.
scanning and magnetic resonance imaging, could assess this. The Hides J, Stokes M, Saide M, Jull G, Cooper D. Evidence of lumbar multifidus muscle
main new contribution of this paper is the data pertaining to wasting ipsilateral to symptoms in patients with acute/subacute low back pain.
Spine 1994;19(2):165–72.
contraction of the multifidus muscle. The measure could be used in Hides JA, Wilson S, Stanton W, McMahon S, Sims K, Richardson C. The effect of
future studies to compare the effectiveness of retraining motor stabilization training on multifidus muscle size among young elite cricketers.
control with and without feedback by ultrasound imaging in Journal of Orthopaedic and Sports Physical Therapy 2008b;38(3):101–8.
International Association for the Study of Pain. Available at: http://www.iasp-pain.
subjects with CLBP. org//AM/Template.cfm?Section¼Home; 2008 [accessed April 1, 2008].
Kader D, Wardlaw D, Smith F. Correlation between the MRI changes in the lumbar
4.2.2. Conclusion multifidus muscles and leg pain. Clinical Radiology 2000;55:145–9.
Kiesel K, Uhl T, Underwood F, Rodd D, Nitz A. Measurement of lumbar multifidus
Patients with CLBP had significantly smaller multifidus muscles
muscle contraction with rehabilitative ultrasound imaging. Manual Therapy
than healthy, asymptomatic subjects at the lowest vertebral level of 2007;12(2):161–6.
the lumbar spine. Patients with CLBP also had greater difficulty Kiesel KB, Uhl TL, Underwood FB, Nitz A. Rehabilitative ultrasound measurement t
of select trunk muscle activation during induced pain. Manual Therapy 2008;
performing a voluntary isometric multifidus contraction at the
13:132–8.
same vertebral level. The results of this study support previous O’Sullivan PB, Twomey LT, Allison GT. Evaluation of specific stabilizing exercise in
findings that the pattern of multifidus muscle atrophy in CLBP the treatment of chronic low back pain with radiologic diagnosis of spondy-
patients is localized rather than generalized but goes further in also lolysis or spondylolisthesis. Spine 1997;22:2959–67.
Pressler JF, Heiss GD, Buford JA, Chidley JV. Between day repeatability and symmetry
ascertaining a reduced ability to voluntarily contract the atrophied of multifidus cross-sectional area measured using ultrasound imaging. Journal of
muscle. These findings lend support to the use of specific muscle Orthopaedic and Sports Physical Therapy 2006;36(1):10–8.
retraining programmes for patients with CLBP. Sihvonen T, Herno A, Paljarvi L, Airaksinen O, Partanen J, Tapaninaho A. Local
denervation atrophy of paraspinal muscles in postoperative failed back
syndrome. Spine 1993;18:575–81.
Acknowledgements Stokes M, Rankin G, Newham DJ. Ultrasound imaging of lumbar multifidus: normal
reference ranges for measurements and practical guidance on the technique.
Manual Therapy 2005;10:116–26.
The authors wish to thank the subjects studied, the staff at the Stokes M, Hides J, Elliott J, Kiesel K, Hodges P. Rehabilitative ultrasound imaging of
UQ/Mater Back Stability Clinic, and its Director, Ms Linda Blackwell, the posterior paraspinal muscles. Journal of Orthopaedic and Sports Physical
for their assistance, and Ms Margot Wilkes, for assistance with this Therapy 2007;37(10):581–95.
Stuge B, Veierød MB, Laerum E, Vøllestad N. The efficacy of a treatment program
manuscript.
focusing on specific stabilizing exercises for pelvic girdle pain after pregnancy:
a two year follow up of a randomized clinical trial. Spine 2004;29(10):
References E197–203.
Van K, Hides JA, Richardson CA. The use of real-time ultrasound imaging for
Barker KL, Shamley DR, Jackson D. Changes in the cross-sectional area of multifidus biofeedback of lumbar multifidus muscle contraction in healthy subjects.
and psoas in patients with unilateral back pain: the relationship to pain and Journal of Orthopaedic and Sports Physical Therapy 2006;36(12):920–5.
disability. Spine 2004;29:E515–9. Wallwork T, Hides J, Stanton W. Intrarater and interrater reliability of assessment
Danneels L, Vanderstraeten G, Cambier D, Witvrouw E, De Cuyper H. CT imaging of of lumbar multifidus muscle thickness using rehabilitative ultrasound
trunk muscles in chronic low back pain patients and healthy control subjects. imaging. Journal of Orthopaedic and Sports Physical Therapy
European Spine Journal 2000;9(4):266–72. 2007;37(10):608–12.