See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/277252736

Prevalence of gluteus medius weakness in people with chronic low back pain
compared to healthy controls

Article  in  European Spine Journal · May 2015

DOI: 10.1007/s00586-015-4027-6 · Source: PubMed

CITATIONS READS

54 2,252

7 authors, including:

Nicholas A Cooper Dennis Bewyer

Saint Ambrose University University of Iowa
6 PUBLICATIONS   226 CITATIONS    11 PUBLICATIONS   197 CITATIONS   

SEE PROFILE SEE PROFILE

Kathleen Sluka
University of Iowa
311 PUBLICATIONS   15,082 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Chemical and pharmacological study of medicinal plants rich in essential oils and their main compounds View project

Development and pharmacological study of pharmaceutical preparations containing cyclodextrins View project

All content following this page was uploaded by Kathleen Sluka on 07 December 2015.

The user has requested enhancement of the downloaded file.

Eur Spine J
DOI 10.1007/s00586-015-4027-6

ORIGINAL ARTICLE

Prevalence of gluteus medius weakness in people with chronic low

back pain compared to healthy controls
Nicholas A. Cooper1,2 • Kelsey M. Scavo1 • Kyle J. Strickland1 • Natti Tipayamongkol1 •

Jeffrey D. Nicholson2 • Dennis C. Bewyer2 • Kathleen A. Sluka1

Received: 13 May 2014 / Revised: 13 May 2015 / Accepted: 13 May 2015

Ó Springer-Verlag Berlin Heidelberg 2015

Abstract test, p \ 0.001). The Trendelenburg sign is more prevalent

Purpose Clinical observation suggests that hip abductor in subjects with LBP than controls (Cochran’s Q,
weakness is common in patients with low back pain (LBP). p \ 0.001). There is more palpation tenderness over the
The purpose of this study is to describe and compare the gluteals, greater trochanter, and paraspinals in people with
prevalence of hip abductor weakness in a clinical popula- low back pain compared to controls (Cochran’s Q,
tion with chronic non-specific LBP and a matched sample p \ 0.001). Hierarchical linear regression, with BMI as a
without LBP. covariate, demonstrated that gluteus medius weakness, low
Methods One hundred fifty subjects with chronic non- back regional tenderness, and male sex were predictive of
specific LBP and a matched cohort of 75 control subjects LBP in this sample.
were recruited. A standardized back and hip physical exam Conclusion Gluteus medius weakness and gluteal muscle
was performed. Specifically tensor fascia lata, gluteus tenderness are common symptoms in people with chronic
medius, and gluteus maximus strength were assessed with non-specific LBP. Future investigations should validate
manual muscle testing. Functional assessment of the hip these findings with quantitative measures as well as in-
abductors was performed with assessment for the presence vestigate the effect of gluteus medius strengthening in
of the Trendelenburg sign. Palpation examination of the people with LBP.
back, gluteal and hip region was performed to try and re-
produce the subject’s pain complaint. Friedman’s test or Keywords Low back pain
Gluteus medius
Hip
Cochran’s Q with post hoc comparisons adjusted for abductor
Muscle pain
Muscle tenderness
Muscular
multiple comparisons was used to compare differences weakness
between healthy controls and people with chronic low back
pain for both the affected and unaffected sides. Mann–
Whitney U was used to compare differences in prevalence Introduction
between groups. Hierarchical linear regression was used to
identify predictors of LBP in this sample. Low back pain (LBP) was recently reported to be the single
Results Gluteus medius is weaker in people with LBP largest cause of disability across the globe [1]. Current
compared to controls or the unaffected side (Friedman’s interdisciplinary practice guidelines show strong evidence
for exercise as an intervention for LBP and thus recom-
mend exercise for management of both acute and chronic
& Nicholas A. Cooper LBP [2, 3]. A variety of exercise interventions have been
[email protected] studied in randomized controlled trials and have been
1 shown to improve pain and disability in people with LBP
Department of Physical Therapy and Rehabilitation Science,
Carver College of Medicine, University of Iowa, Iowa City, [4, 5]. Despite this support, it remains unclear which ex-
IA 52242, USA ercise interventions are optimal in people with chronic LBP
2
Department of Rehabilitation Therapies, University of Iowa [5]. Current physical therapy guidelines suggest several
Hospitals and Clinics, Iowa City, IA 52242, USA possible exercise treatment strategies depending on the

123
 Eur Spine J

patient’s presentation [6]. Most patients with long-standing hundred fifty-one subjects were approached and agreed to
LBP are matched to exercise treatment based on the phy- participate with one subject withdrawing after participa-
sical therapy evaluation. Specifically, exercise interven- tion. After recruitment of the LPB group, an age and sex-
tions provided by physical therapists typically focus on the matched cohort of 75 control subjects was recruited. Sub-
abdominal and lumbar musculature strength or directional jects were questioned about their personal history with
preference exercises [6–8]. LBP: only subjects reporting no to having current LBP and
An alternative to these interventions may be to focus on no to having a history of more than 3 months of LBP were
the hip abductor musculature. Simons and Travell describe included. Two subjects were excluded for histories of more
myofascial pain from the gluteus medius muscle as a than 3 months of LBP. The University of Iowa Institutional
common component of LBP [9]. Subsequently, Njoo and Review Board approved this study and informed consent
van der Does reported a higher prevalence of gluteus was obtained from all subjects.
medius myofascial trigger points in people with LBP [10].
We also suggested gluteus medius weakness is associated Screening examination
myofascial pain and trochanteric bursitis is a common
clinical presentation in people with LBP [11]. In addition All subjects were screened for exclusionary diagnoses with
to myofascial pain, weakness of hip abductors has been a standardized history and physical examination. This in-
described in LBP when compared to healthy controls [12– cluded questions screening for lower extremity paresthesia
14]. Further, asymmetry in hip abductor strength has been and weakness, bowel and bladder dysfunction, pre-
correlated with increased likelihood to seek care for LBP in dominant lower extremity pain with standing and walking,
collegiate athletes and we reported gluteus medius weak- history of trauma, presence of systemic illness, weight loss,
ness was associated with onset of LBP during pregnancy and predominant night pain. The physical examination
[15, 16]. More recently Nelson-Wong and colleagues re- screening included assessment for reflex asymmetry, my-
ported that subjects who developed LBP during an ex- otomal weakness, sensory disturbance, straight leg raise,
perimental standing task had a different recruitment pattern and groin pain with hip internal rotation. Subjects were
of gluteus medius muscle compared to those who did not excluded if screening was suggestive of specific pathology.
develop LBP [17]. Although these findings are suggestive Potential control subjects were screened identically and
of hip abductors playing a role in LBP, it is unclear what additionally were excluded if they presented with either
proportion of the population with LBP presents with hip acute or chronic low back pain.
abductor weakness and associated symptoms when com-
pared to healthy controls as well as which hip abductor Muscle strength
muscles are weak.
In this study, we quantified the prevalence of hip muscle Gluteus medius, TFL, and gluteus maximus manual muscle
weakness and tenderness of the hip and low back in people tests (MMTs) were performed using break tests as de-
with chronic non-specific LBP. We hypothesized that scribed by Hislop and Montgomery [19]. Gluteus medius
gluteus medius weakness and tenderness occurs in the strength was tested by placing subject in side-lying and
majority of people with non-specific chronic LBP com- having the subject abduct and slightly extend the hip while
pared to people without LBP. keeping the pelvis rotated slightly forward. Resistance was
applied at the ankle. TFL strength was tested by position-
ing the subject in side-lying with the limb to be tested
Methods flexed at the hip. The hip was abducted in this flexed po-
sition and resistance was applied at the ankle. Gluteus
Subjects maximus strength was tested positioning the subject in
prone with the knee flexed, then hip was extended with the
Two groups of subjects were recruited. One hundred fifty knee remaining flexed and resistance was applied at the
people seeking care for LBP lasting longer than 3 months posterior thigh just above the knee. MMTs were scored
at the University of Iowa Spine Center Physical Therapy using the criteria defined by Hislop and Montgomery [19].
Clinic were serially recruited at time of presentation to the If the subject was able to resist maximal resistance they
clinic. Subjects were recruited if they had non-specific were scored 5/5, if they broke against resistance: 4/5; if
LBP, defined as pain anywhere from the inferior rib margin they were unable to tolerate resistance, but could move
to the gluteal fold, for more than 3 months [18]. Subjects against gravity: 3/5; if they could move the limb when
with a defined etiology, including radiculopathy, neuro- positioned to minimize the effect of gravity: 2/5; palpable
genic claudication, fracture, primary or secondary spinal contraction, but no movement: 1/5, and no palpable ac-
tumors, or other specific pathology, were excluded. One tivity: 0/5.

123
Eur Spine J

Trendelenburg sign t tests. Inter-rater reliability of MMT, Trendelenburg sign,

and palpation tenderness was assessed with a two-way
The Trendelenburg sign was assessed as a functional random effects model. Manual muscle testing results be-
measure of gluteus medius strength [20]. This was per- tween groups and side were assessed with Friedman’s test
formed as outlined by Hardcastle and Nade [21]. Subjects with post hoc comparisons adjusted for multiple compar-
stood with the examiner behind them, observing the iliac isons. The presence of the Trendelenburg sign and palpa-
crests visually and with palpation, and were instructed to tion tenderness were assessed between groups and side
lift one foot off the ground by flexing the hip. The sign was using Cochran’s Q corrected for multiple comparisons.
considered absent if the subject was able to maintain the A Mann–Whitney U was used to compare differences in
pelvis in neutral or with the non-stance side elevated and prevalence between groups. A receiver operating charac-
present if the subject was unable to maintain the pelvis teristic (ROC) curve was generated to determine the ap-
level or had to shift the trunk to keep the pelvis level. propriate cut off for MMT scores for determining weakness
in the total sample population. A hierarchical linear re-
Tenderness gression was performed to identify predictors of the pres-
ence of LBP in the total sample population. Any
The gluteals, greater trochanters, lumbar paraspinals, and demographic differences between samples were treated as
piriformis were palpated for tenderness bilaterally. Gluteus covariates. Potential predictors included age, sex, BMI,
medius was palpated from its distal insertion at the greater presence of gluteus medius weakness, low back regional
trochanter over the muscle belly toward the posterior su- tenderness to palpation, and the presence of Trendelenburg
perior iliac spine (PSIS) and then over its proximal at- sign. Regression was performed in both forward and
tachment along the ilium just inferior to the iliac crest. backward methods using p \ 0.05 as entry criterion and
Gluteus maximus was palpated at its origin along the pos- p [ 0.1 as exit criterion. Significance was set at p = 0.01
terior ilium and lateral sacrum, then over the muscle belly to for all tests. SPSS 21 was used for all analyses.
its distal insertion at the iliotibial band inferior to the greater
trochanter. The greater trochanters were palpated most
laterally initially and then posteriorly and superiorly to the Results
apex of the trochanter. The lumbar paraspinals were pal-
pated from just medial to the PSIS superiorly to the thorax. One hundred fifty subjects with chronic LBP and 75 age
Palpation of the piriformis was attempted from its lateral and sex-matched control subjects were recruited and en-
insertion at the greater trochanter, over the muscle belly, rolled. Characteristics of both groups are presented in
toward its origin on the sacrum. Tenderness was defined as Table 1. Subjects with chronic LBP had a higher BMI.
reproduction of the subject’s pain complaint when using Eighty-four subjects with LBP had unilateral symptoms
enough pressure to blanch the examiner’s nail. and 66 had bilateral symptoms, totaling 216 affected and
84 unaffected sides. Interclass correlation coefficient for
Inter-rater reliability MMT was 0.597 (0.331–0.774, 95 % CI). There was per-
fect reliability among Trendelenburg sign and muscle
Inter-rater reliability was assessed between all four asses- tenderness examination.
sors. Assessors had 1–40 years in performing these There were significant main effects for MMT of the
assessments. All assessors independently and in random gluteus medius, TFL, and gluteus maximus (Friedman’s
order assessed a series of six female pilot subjects with a test, all p \ 0.001) (Fig. 1). A significant decrease in glu-
mean age of 30 years, three with LBP and three without teus medius strength was observed for the affected side
LBP, using the above described MMT, Trendelenburg sign, (MMT grade ± SD, 3.35 ± 0.73) compared to the unaf-
and muscle tenderness examination. The examiners were fected side (4.56 ± 0.66, p \ 0.001) or control group
blinded to the presence of LBP in these subjects. (4.46 ± 0.50, p \ 0.001). TFL strength was significantly
greater on the unaffected side (4.93 ± 0.26) compared to
Data analysis controls (4.48 ± 0.50, p \ 0.001), but not the affected side
(4.81 ± 0.44). There were no significant differences in
Data from subjects with chronic LBP were divided into gluteus maximus strength. There was a significant main
affected and unaffected sides based on the location of their effect for the Trendelenburg sign between groups
symptoms. Subjects with bilateral complaints were treated (Cochran’s Q, p \ 0.001). The Trendelenburg sign was
as both sides being affected sides. The percentage of sub- more frequently present on the affected side (54.2 %)
jects with each symptom was calculated. Age, height, compared to the unaffected side (7.1 %, p \ 0.001) or
weight, and BMI were compared with independent samples controls (9.7 %, p \ 0.001) (Fig. 1d).

123
 Eur Spine J

Table 1 Subject characteristics

Chronic LBP (n = 150) Control (n = 75) t test
are mean ± standard deviation
Age (years) 41.4 ± 13.0 40.7 ± 13.9 0.329, p = 0.743
Sex (% female) 64.7 65.3
Height (cm) 169.4 ± 11.4 168.2 ± 9.4 0.801, p = 0.424
Weight (kg) 84.9 ± 22.2 73.2 ± 21.2 3.623, p \ 0.001
BMI (kg/m2) 29.6 ± 7.2 25.8 ± 7.0 3.543, p \ 0.001
Only weight and BMI were significantly different between groups

A 5 B 5
+
G. Medius MMT (0-5)

4 4
*

TFL MMT (0-5)

3 3

2 2

1 1

0 0
Affected Unaffected Control Affected Unaffected Control

C 5 D 100
90
(+) Trendelenburg (%)

80
G. Maximus MMT (0-5)

4
70
3 60
50
*
2 40
30
1 20
10
0 0
Affected Unaffected Control Affected Unaffected Control

Fig. 1 Gluteus medius strength is significantly less on the affected (?p \ 0.001). Trendelenburg sign is significantly more prevalent on
side compared to the unaffected side or controls (*p \ 0.001). TFL the affected side compared to both the unaffected side and controls
strength is greater on the unaffected side compared to controls (*p \ 0.001)

There was a significant difference for the presence of (0.7 %, p \ 0.001), as well as being different between the
palpation tenderness between groups (Mann–Whitney U, unaffected side and controls (p = 0.006). There was no
p \ 0.001) (Fig. 2a). There were significant main effects significant difference in piriformis tenderness.
for palpation tenderness over the gluteals, greater tro- The ROC curve demonstrated that an MMT score of B3/
chanter, and lumbar paraspinals (Cochran’s Q, p \ 0.001) 5 was the optimal cut point for using gluteus medius
(Fig. 2b–d). Gluteal tenderness was more prevalent on the weakness to determine the presence of LBP from the total
affected side (68.1 %) compared to the unaffected side sample (Fig. 3). The presence of LBP was correlated with
(4.8 %, p \ 0.001) or controls (11.2 %, p \ 0.001). higher BMI, gluteus medius weakness, low back tender-
Similarly, there was more frequent tenderness over the ness, and a positive Trendelenburg sign (Table 2). Subse-
greater trochanter on the affected side (44.9 %) compared quent hierarchical linear regression found that higher BMI,
to the unaffected side (6.0 %, p \ 0.001) or controls gluteus medius weakness, low back regional tenderness,
(6.0 %, p \ 0.001). Lumbar paraspinal tenderness was and male sex as predictors of LBP across this sample
more prevalent on the affected side (53.2 %) compared to (Table 3). The strongest contributor was gluteus medius
the unaffected side (23.8 %, p \ 0.001) or the controls weakness (DR2 = 0.461, p \ 0.001) while the other

123
Eur Spine J

A 100
B 100

TTP over Gluteals (%)

90 * 90

Low Back TTP (%)

80 80
70 70 *
60 60
50 50
40 40
30 30
20 20
10 10
0 0
LBP Control Affected Unaffected Control
C D
100 100

TTP over Paraspinals (%)

TTP over Trochanter (%)

90 90
80 80
70 70
60 60
50 50 *
40
* 40
30 30 +
20 20
10 10
0 0
Affected Unaffected Control Affected Unaffected Control

Fig. 2 Low back tenderness is more prevalent in subjects with LBP both the unaffected side and controls (*p \ 0.001). There is more
(*p \ 0.001). Tenderness is more prevalent over the gluteals, greater paraspinal tenderness on the unaffected side compared to controls
trochanter, and lumbar paraspinals on the affected side compared to (?p = 0.001)

1.0 However, only the gluteus medius weakness was a sig-

3/5 or less MMT
0.9 nificant predictor of the presence of LBP when compared
0.8 to the controls. Future studies will need to confirm this
0.7 subgroup in other settings and determine if treatment of the
gluteus medius weakness has a positive effect on
Sensitivity

0.6
4/5 or less MMT symptoms.
0.5
Our data agree with prior studies showing relative hip
0.4
abductor weakness in subjects with LBP compared to
0.3
controls [13, 14]. Both of these studies look at hip ab-
0.2
duction in the frontal plane with contributions from the
0.1 TFL and gluteus medius. Neither study assessed the com-
0.0 posite abduction and extension that is performed by the
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
1-Specificity
gluteus medius with minimal TFL contributions [19]. The
current study was specifically designed to separate out the
Fig. 3 Using a cutoff of 3/5 or less on gluteus medius strength testing contribution of the TFL from the gluteus medius by at-
was most accurate in the assessment of LBP in this sample tempting to evaluate each individually. Our results
demonstrated no difference in TFL strength between con-
variables explained substantially less of the variability trols and the affected side of subjects with LBP, but sig-
(DR2 \ 0.1, p B 0.001). nificant weakness in the gluteus medius in subjects with
LBP. Thus, the current study suggests gluteus medius
muscle weakness contributes to the presentation of chronic
Discussion non-specific LBP. Reproduction of these results with a
quantitative strength assessment such as dynamometry
The current study identifies a sub-population of patients could be used to confirm the current results. Further the
with chronic non-specific LBP with signs of hip abductor effects of treatment of the gluteus medius muscle weakness
dysfunction: significant gluteus medius weakness (B3/5 on LBP itself would further help determine the relevance of
MMT), gluteal tenderness, and the Trendelenburg sign. the weakness to the pain itself. Additionally, there may be

123
 Eur Spine J

Table 2 The presence of LBP was correlated with higher BMI, gluteus medius muscle weakness, low back regional tenderness and a positive
Trendelenburg sign
LBP Age Sex BMI G medius weak Low back tender

Age 0.021, p = 0.754

Sex 0.083, p = 0.221 -0.072, p = 0.291
BMI 0.281, p \ 0.001 0.0199, p = 0.003 0.051, p = 0.455
G medius weak 0.723, p \ 0.001 0.004, p = 0.959 -0.084, p = 0.216 0.213, p = 0.002
Low back tender 0.672, 0 \ 0.001 0.052, p = 0.448 -0.078, p = 0.252 0.298, p \ 0.001 0.667, p \ 0.001
Trendelenburg 0.452, p \ 0.001 0.082, p = 0.230 -0.027, p = 0.688 0.075, p = 0.269 0.568, p \ 0.001 0.505, p \ 0.001

Table 3 Hierarchical linear

Beta Standard error t p Partial R2 Model R2
regression with BMI as a
covariate Constant 0.061 0.083 0.739 0.461 0.615
BMI 0.005 0.003 1.527 0.128 0.079
G medius weakness 0.465 0.053 8.763 \0.001 0.461
Low back tender 0.317 0.057 5.601 \0.001 0.054
Male sex 0.146 0.042 3.445 0.001 0.022
Gluteus medius weakness is the strongest predictor of LBP. Low back tenderness and male sex also were
predictive of LBP

some amount of gluteus medius muscle weakness in the system [23]. There does not exist a comprehensive classi-
population in general. A large proportion of our control fication system that directs intervention and successfully
population scored only 4/5 on MMT of gluteus medius predicts outcome. One of the difficulties in directing ex-
muscle and thus this could be a potential risk factor for ercise treatment in chronic LBP may be the broad effect of
development of non-specific LBP. any exercise intervention.
Others have also suggested that gluteus medius dys- Most exercise interventions in chronic LBP populations
function plays a role in LBP. Simons and Travell described are effective [4]. The most recent Cochrane meta-analysis
gluteus medius muscle referred pain as a component of of exercise in LBP found that individually prescribed ex-
LBP [9]. Nadler and colleagues demonstrated a higher ercise interventions that included strengthening and stabi-
likelihood of onset of LBP in female athletes with hip lization exercises were most common in chronic LBP
abductor strength differences between sides [16]. Nelson- populations [24]. These exercise interventions were con-
Wong and colleagues demonstrated gluteus medius co-ac- cluded to be effective in improving pain and function in
tivation as a predictor of onset LBP during an experimental chronic LBP [24]. This finding was reiterated in the clinical
standing task [17]. Together these data suggest that gluteus practice guidelines from the American Pain Society and
medius muscle may play a significant role in chronic LBP. American College of Physicians [3]. However, they do not
However, it is unclear if initial gluteus medius muscle differentiate between choices of exercise intervention [3].
weakness is a cause or consequence of LBP as well as how One of the chief reasons for the paucity of advice regarding
to manage this observed dysfunction. specific exercise selection is the poor description of exer-
Current physical therapy management of patients with cise interventions in the literature. Future experiments
LBP is guided by a treatment-based classification system should examine specific populations with well-described
that attempts to match subgroups of patients with the in- targeted exercise programs [24]. We believe patients with
terventions that lead to the best outcomes [6]. Much of this chronic non-specific LBP who present with gluteus medius
work has focused on acute LBP; patients with chronic LBP weakness and associated tenderness may represent a
do not readily fit into this classification system [22]. A treatment subgroup that could benefit from targeted gluteus
recent review of classification systems for chronic LBP medius strengthening. A preliminary study of ten subjects
found strong evidence to support the reliability of only two with non-specific LBP treated with 3 weeks of hip abductor
systems: the McKenzie and the movement impairment strengthening reduced pain by 48 %, but was not statisti-
classification systems [23]. They also report some evidence cally significant [14]. Future clinical studies will need to
to support the effectiveness of the McKenzie classification confirm that gluteus medius strengthening produces

123
Eur Spine J

superior results than other forms of exercise in this sub- Acknowledgments This work was supported by the Department of
group of patients with chronic LBP. Rehabilitation Therapies at the University of Iowa Hospitals and
Clinics; the Department of Physical Therapy and Rehabilitation Sci-
The current study demonstrated a statistical difference in ence at the Carver College of Medicine at the University of Iowa; and
TFL strength between the unaffected side of subjects with the National Center for Research Resources and the National Center
LBP and control subjects. The assessed strength values are for Advancing Translational Sciences, National Institutes of Health,
extremely close: 4.48 in the controls and 4.93 in the sub- through grant UL1RR024979.
jects with LBP on the unaffected side. Although this dif- Conflict of interest None.
ference reached statistical significance we believe this does
not represent a clinically significant difference, especially
given the lack of objective criteria for MMT grades above
References
3/5 [19]. Use of dynamometry may better evaluate this
apparent difference in muscle strength. 1. Hoy D, March L, Brooks P, Blyth F, Woolf A, Bain C, Williams
There are several limitations in this study. The clinical G, Smith E, Vos T, Barendregt J, Murray C, Burstein R, Buch-
sample was recruited from one specialty physical therapy binder R (2014) The global burden of low back pain: estimates
from the Global Burden of Disease 2010 study. Ann Rheum Dis
clinic in a major university hospital. The population re-
73:968–974
ferred to our clinic may be uniquely different than those 2. Chou R, Qaseem A, Snow V, Casey D, Cross JT, Shekelle P,
seen in a general practice. Therefore, future studies should Owens DK (2007) Diagnosis and treatment of low back pain: a
confirm this distribution of subjects across other practice joint clinical practice guideline from the American college of
physicians and the american pain society. Ann Intern Med
settings. Although our controls were matched by age and
147:478–491
sex to the clinical sample, they had a lower BMI. It is 3. Chou R, Huffman LH (2007) Nonpharmacologic therapies for
possible that our control subjects were healthier or more acute and chronic low back pain: a review of the evidence for an
physically active and thus stronger. The assessments we American pain society/American college of physicians clinical
practice guideline. Ann Intern Med 147:492–504
used are performed routinely in clinical settings; however,
4. van Middelkoop M, Rubinstein SM, Kuijpers T, Verhagen AP,
they have limited objectivity. The high inter-rater re- Ostelo R, Koes BW, van Tulder MW (2011) A systematic review
liability for tenderness and Trendelenburg sign may rep- on the effectiveness of physical and rehabilitation interventions
resent extensive training of our examiners prior to data for chronic non-specific low back pain. Eur Spine J 20:19–39
5. van Middelkoop M, Rubinstein SM, Verhagen AP, Ostelo RW,
collection. While inter-rater reliability of MMT is low,
Koes BW, van Tulder MW (2010) Exercise therapy for chronic
this is similar to prior reports [25]. Further, we ac- nonspecific low-back pain. Best Pract Res Clin Rheumatol
knowledge that MMT used clinically is an indirect mea- 24:193–204
sure of muscle function and not as strong as dynamometry 6. Delitto A, George SZ, Van Dillen LR, Whitman JM, Sowa G,
Shekelle P, Denninger TR, Godges JJ (2012) Low back pain.
[26]. Using palpation, a commonly used clinical measure,
J Orthop Sports Phys Ther 42:A1–57
is similarly less specific than using pressure pain thresh- 7. Standaert CJ, Weinstein SM, Rumpeltes J (2008) Evidence-in-
olds, which could be confirmed with pressure algometry. formed management of chronic low back pain with lumbar sta-
Future studies should utilize quantitative assessments bilization exercises. Spine J 8:114–120
8. Mayer J, Mooney V, Dagenais S (2008) Evidence-informed
such as pressure algometry to measure pain thresholds,
management of chronic low back pain with lumbar extensor
dynamometry to measure muscle strength, and elec- strengthening exercises. Spine J 8:96–113
tromyography to examine muscle activation patterns to 9. Simons DG, Travell JG (1983) Myofascial origins of low back
directly compare and confirm the clinical measures of pain. 3. Pelvic and lower extremity muscles. Postgrad Med
73(99–105):108
palpation tenderness and MMT. Future studies should also
10. Njoo KH, Van der Does E (1994) The occurrence and inter-rater
determine if strengthening of the gluteus medius muscle reliability of myofascial trigger points in the quadratus lumborum
will alleviate LBP symptoms and which of these measures and gluteus medius: a prospective study in non-specific low back
individually or in combination predicts outcomes to pain patients and controls in general practice. Pain 58:317–323
11. Bewyer DC, Bewyer KJ (2003) Rationale for treatment of hip
treatment.
abductor pain syndrome. Iowa Orthop J 23:57–60
12. Nourbakhsh MR, Arab AM (2002) Relationship between me-
chanical factors and incidence of low back pain. J Orthop Sports
Conclusion Phys Ther 32:447–460
13. Arab A, Nourbakhsh M (2010) The relationship between hip
abductor muscle strength and iliotibial band tightness in indi-
Gluteus medius muscle weakness and associated tender- viduals with low back pain. Chiropr Osteopat 18:1
ness is a common presentation in people with chronic non- 14. Kendall KD, Schmidt C, Ferber R (2010) The relationship be-
specific LBP. Focusing on assessment and treatment of tween hip-abductor strength and the magnitude of pelvic drop in
patients with low back pain. J Sport Rehabil 19:422–435
gluteus medius muscle dysfunction may allow for better
15. Bewyer KJ, Bewyer DC, Messenger D, Kennedy CM (2009) Pilot
clinical decision-making and better treatment outcomes for data: association between gluteus medius weakness and low back
people with LBP. pain during pregnancy. Iowa Orthop J 29:97–99

123
 Eur Spine J

16. Nadler SF, Malanga GA, Bartoli LA, Feinberg JH, Prybicien M, 21. Hardcastle P, Nade S (1985) The significance of the Trendelen-
Deprince M (2002) Hip muscle imbalance and low back pain in burg test. J Bone Joint Surg Br 67:741–746
athletes: influence of core strengthening. Med Sci Sports Exerc 22. Apeldoorn AT, Ostelo RW, van Helvoirt H, Fritz JM, Knol DL,
34:9–16 van Tulder MW, de Vet HC (2012) A randomized controlled trial
17. Nelson-Wong E, Gregory DE, Winter DA, Callaghan JP (2008) on the effectiveness of a classification-based system for subacute
Gluteus medius muscle activation patterns as a predictor of low and chronic low back pain. Spine 37:1347–1356
back pain during standing. Clin Biomech 23:545–553 23. Fairbank J, Gwilym SE, France JC, Daffner SD, Dettori J,
18. Dionne CE, Dunn KM, Croft PR, Nachemson AL, Buchbinder R, Hermsmeyer J, Andersson G (2011) The role of classification of
Walker BF, Wyatt M, Cassidy JD, Rossignol M, Leboeuf-Yde C, chronic low back pain. Spine 36:S19–S42
Hartvigsen J, Leino-Arjas P, Latza U, Reis S, Gil Del Real MT, 24. Hayden JA, van Tulder MW, Malmivaara A, Koes BW (2005)
Kovacs FM, Oberg B, Cedraschi C, Bouter LM, Koes BW, Pi- Exercise therapy for treatment of non-specific low back pain.
cavet HS, van Tulder MW, Burton K, Foster NE, Macfarlane GJ, Cochrane Database Syst Rev 20(3):CD000335
Thomas E, Underwood M, Waddell G, Shekelle P, Volinn E, Von 25. Frese E, Brown M, Norton BJ (1987) Clinical reliability of
Korff M (2008) A consensus approach toward the standardization manual muscle testing. Middle trapezius and gluteus medius
of back pain definitions for use in prevalence studies. Spine muscles. Phys Ther 67:1072–1076
33:95–103 26. Krause DA, Neuger MD, Lambert KA, Johnson AE, DeVinny
19. Hislop HJ, Montgomery J (2002) Daniels and Worthingham’s HA, Hollman JH (2014) Effects of examiner strength on re-
muscle testing : techniques of manual examination. Saunders, liability of hip-strength testing using a handheld dynamometer.
Philadelphia J Sport Rehabil 23:56–64
20. Trendelenburg F (1998) Trendelenburg’s test: 1895. Clin Orthop
Relat Res 355:3–7

123

View publication stats