Aladro Et Al., (2012) - Pilates-Based Exercise For Persistent, Non-Specific Low Back Pain and Associated Functional Disability. A Meta-Analysis With Meta-Regression

This article appeared in a journal published by Elsevier.
The attached
copy is furnished to the author for internal non-commercial research
and education use, including for instruction at the authors institution
and sharing with colleagues.
Other uses, including reproduction and distribution, or selling or
licensing copies, or posting to personal, institutional or third party
websites are prohibited.
In most cases authors are permitted to post their version of the
article (e.g. in Word or Tex form) to their personal website or
institutional repository. Authors requiring further information
regarding Elseviers archiving and manuscript policies are
encouraged to visit:
http://www.elsevier.com/copyright
Author's personal copy

Journal of Bodywork & Movement Therapies (2013) 17, 125e136
Available online at www.sciencedirect.com
PREVENTION & REHABILITATION e PILATES EXERCISE: META ANALYSIS
Pilates-based exercise for persistent, non-specific

low back pain and associated functional disability:
A meta-analysis with meta-regression
n R. Aladro-Gonzalvo, Lic a,*, Gerardo A. Araya-Vargas, M.Sc a,
Aria
Mriam Machado-Daz, MD b, Walter Salazar-Rojas, Ph.D a
a
b
School of Physical Education and Sports, University of Costa Rica, San Jose, Costa Rica
Medicine Sport Center of Cienfuegos, Cuba
Received 27 March 2012; received in revised form 29 July 2012; accepted 17 August 2012
KEYWORDS
Pilates;
Systematic review;
Exercise therapy;
Complementary
therapy;
Lumbar spine;
Rehabilitation;
Meta-regression
Summary Objective: The purposes of this study were to systematically review and apply
regression analysis to randomised controlled trials [RCTs] that evaluated the effectiveness
of Pilates exercise in improving persistent, non-specific low back pain and functional disability.
Methods: Electronic databases were searched from January 1950 to March 2011. Articles were
eligible for inclusion if they were RCTs comparing Pilates exercise with a placebo treatment
[PT], minimal intervention [MI] or another physiotherapeutic treatment [APT].
Results: Nine trials were included. Pilates was moderately superior to APT (pooled Effect Size
[ES] weighted Z 0.55, 95% confidence interval [CI] Z 0.08 to 1.03) in reducing disability
but not for pain relief. Pilates provided moderate to superior pain relief compared to MI
(pooled ES weighted Z 0.44, 95% CI Z 0.09 to 0.80) and a similar decrease in disability.
The statistical model used did not detect any predictor variable.
Conclusions: Due to the presence of co-interventions and the low methodological quality of
some studies, these conclusions should be interpreted with caution.
Crown Copyright 2012 Published by Elsevier Ltd. All rights reserved.
Introduction
Low back pain [LBP], with or without leg pain, has been one
of the main burdens to public health for years in the
Western industrialised world, with an abundance of records
about its frequency, recurrence, treatment and cost

(Anderson, 1999; Hart et al., 1995; Luo et al., 2004).
Pain and fear-avoidance often result in activity limitations, (e.g., walking, squatting, lifting, prolonged sitting,
reaching and twisting), participation restriction (e.g., work,
* Corresponding author. 44th Ave., Number 3720, Cienfuegos, Cuba. Tel.: 53 43523650.
E-mail address: aladro80@fastmail.es (A.R. Aladro-Gonzalvo).
1360-8592/$ - see front matter Crown Copyright 2012 Published by Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.jbmt.2012.08.003
journal homepage: www.elsevier.com/jbmt
126
recreation, family and community) and functional disability
[FD] (Asghari and Nicholas, 2001; Woby et al., 2004; Mannion
et al., 2001). Additionally research has found that perceived
pain scores demonstrated a strong correlation with disability
measures (Lackner et al., 1996; Waddell et al., 1993).
Treatment options for LBP can be placed into two categories, active and passive (Kirkaldy-Willis and Bernard, 1999).
One increasingly common exercise regimen suggested for
patients with LBP is Pilates-based therapeutic exercise,
adapted and simplified from the traditional and modern
Pilates Method, which usually is defined as a comprehensive
mindebody conditioning method with the main goals of efficient movements, core stability and enhanced performance
(Akuthota and Nadler, 2004; Muscolino and Cipriani, 2004).
Pilates-based therapeutic exercise has been introduced
in the physical therapy community in recent years to
improve rehabilitation programs (Bryan and Hawson, 2003).
Return to functional activities has also been a primary
objective since the method encourages movement earlier
in the treatment process by providing the necessary assistance (Anderson and Spector, 2000).
Several studies have aimed at documenting the benefits
of this modified method in decreasing LBP (Maher, 2004;
Sorosky et al., 2008; La Touche et al., 2008). Recently two
systematic reviews proposed that there is some evidence
supporting the effectiveness of Pilates-based exercises in
the management of LBP. Posadzki et al. (2011) used
a simple descriptive approach to summarise their results,
alerting that no definite conclusions can be drawn. Lim
et al. (2011) used a meta-analytical approach and added
useful information about the magnitude of the effect of
Pilates on pain relief and disability when compared to other
forms of exercise or minimal intervention.
Despite the aforementioned attempts to summarise and
integrate the results of selected trials, up till now,
a regression analysis for predictor variables on treatment
effectiveness has not been applied.
Thus, the objectives of this study were to systematically
review and apply regression analysis to the RCTs that
evaluated the effectiveness of Pilates exercise aimed at
improving persistent, non-specific LBP and associated FD,
and to provide practical and clinically precise information.
Methods
Data sources and searches
A computerised electronic search was performed to identify relevant articles. The search was conducted on MEDLINE (1950eMarch 2011); CINAHL (Cumulative Index to
Nursing and Allied Health Literature) (1982eMarch 2011);
SPORTDISCUS (1975eMarch 2011); ProQuest Dissertations &
Theses (1980eMarch 2011); ScienceDirect (1990eMarch
2011) and Scholar Google. The terms used were Pilates
AND low back pain AND randomised controlled trial OR
systematic review and Pilates AND low back functional
disability AND randomised controlled trial OR systematic
review. Key words relating to the domains of randomised
controlled trials and back pain were used, as recommended
by the Cochrane Back Review Group [CBRG] (Bombardier
et al., 2006).
A.R. Aladro-Gonzalvo et al.

One reviewer (MMD) screened search results for potentially eligible studies, and 2 reviewers (MMD, AAG) independently reviewed articles for eligibility. A third
independent reviewer (GAV) resolved any disagreement
about inclusion or exclusion of trials.
Study selection
The reviewers followed a selection protocol, developed prior
to the beginning of the review that included a checklist for
inclusion and exclusion criterion. Articles were eligible for
inclusion if they were a) RCTs comparing Pilates exercise
with a PT, MI or APT; b) studies carried out on individuals of
all age groups and sex with LBP and FD; c) RCTs reporting
that a criterion for entry was persistent non-specific LBP
(with or without leg pain) of at least 6 weeks duration (not
attributable to any specific disease) or recurrent LBP with at
least 2 painful incidences per year; d) RCTs describing any
restriction or loss of ability associated with LBP, if they reported that a criterion for entry was functional disability; e)
RCTs that reported one of the following outcome measures:
pain or disability; f) RCTs that reported the mean outcome
(e.g., means, standard deviations and sample sizes); and g)
studies that were not restricted to any specific language.
Articles were excluded if they were a) studies that were
not RCTs; b) RCTs that did not provide sufficient information for the calculation of effect size (i.e., means, standard
deviations and sample sizes for the comparison between
groups); and c) RCTs with inconsistent or internal discrepancies in data.
This systematic review followed the recommendations
of the PRISMA statement (Moher et al., 2009).
Data extraction and quality assessment

The first analysis was based on the understanding of the
information provided by the title and abstract. The articles
selected from the first analysis were studied in depth using
the full text in the data extraction phase. Two independent
reviewers (MMD, AAG) extracted data from each study that
was included using a standardised extraction form. Mean
scores, standard deviations, and sample sizes were
extracted from the trials. When there was insufficient
information about outcomes to allow data analysis, the
authors of the study were contacted.
The methodological quality of the trials was assessed
using the PEDro scale (Maher et al., 2003) with scores
extracted from the PEDro database. The PEDro Scale
includes 11 items that, overall, aim to evaluate four
fundamental methodological aspects of a study such as the
random process, the blinding technique, group comparison,
and the data analysis process. The reliability of this scale
was evaluated with acceptable results in intraclass correlation coefficients (ICC) Z 0.56 (95% CI Z 0.47e0.65) for
ratings by individuals, and ICC for consensus ratings Z 0.68
(95% CI Z 0.57e0.76) (Maher et al., 2003).
Assessment of the quality of trials in the PEDro database
was performed by 2 trained independent raters, and
disagreements were resolved by a third rater (Sherrington
et al., 2000).

Pilates-based exercise for persistent low back pain
The data were grouped into two treatment contrasts: (1)

Pilates-based exercise versus minimal intervention (no
intervention, general practitioner care, education), and (2)
Pilates-based exercise versus another physiotherapeutic
treatment.
The calculation of the effect size was used as an estimator of treatment effects (Hedges, 1981). Several
different methods have been proposed for the calculation
of effect size. In this analysis, Thomas and Frenchs formula
was used for the calculation of ES (Thomas and French,
1986). This formula is: ES Z [(PostestEG PostestCG)/
SDPostest CG], where EG Z experimental group (i.e., Pilates
group), CG Z comparison group, SD Z standard deviation.
Because studies with small samples sizes may have
a biased treatment effect, each ES was then multiplied by a
correction factor designed to yield an unbiased
estimate of effect size (Thomas and French, 1986). This
correction factor is: EScorrection Z ES C, where
C Z 1 [3/(4 M 1)], and where M Z nCG 1 (nCG:
number of participants in the comparison group).
Supplemental procedures (e.g., formulas for estimating
pooled ES weighted and coding the studies) are described
as attachment material.
Testing homogeneity
To determine whether all ESs of the studies reviewed were
homogeneous, and whether they represented a similar
measure of treatment effectiveness, a test for homogeneity
was conducted. The homogeneity statistic (H ) was used to
test the null hypothesis, H0: ES1 Z ES2 Z . Z ESi. When
the null hypothesis was not rejected, all ESs were similar
and represented a similar measure of treatmenteffectiveness, and the opposite. For this analysis the
following
P formula2 (Thomas et Pal., 2005) was used:
HZ
Division ES =Variance Division ES=Variance2 =
P
Variance Reciprocal. Under the null hypothesis, H has
a chi-square distribution with N 1 degrees or freedom,
where N equals the number of ESs.
Weighted regression
Regression procedures were conducted to determine the
presence of predictor variables on the treatment effectiveness (Thomas and French, 1986). This technique was
realised for ESs significantly different from zero. Each ES
was weighted by the reciprocal of its variance in the
weighted regression technique. The sum of square total for
the regression was equivalent to the homogeneity statistic,
H. A non-significant test of model specification indicated
that the ESs did not deviate substantially from the regression model, which suggested that several features prior
coding did not have any influence on the treatment effectiveness, and the opposite (Thomas et al., 2005).
Accounting for publication bias
We estimated the number of studies sitting around in file
drawers using the following formula: Ko Z [(k(dpooled ES dES
trivial))/dES trivial], where Ko is the number of studies needed
to produce a trivial ES, K is the number of studies in the
meta-analysis, dpooled ES is the mean of all the ESs in the
meta-analysis, and dES trivial is the estimate of a trivial ES

(e.g., 0.15) (Hedges and Olkin, 1985).
Statistical analysis was performed using SPSS, version
15.0.
Results
The initial electronic database search resulted in a total of
19 articles potentially eligible based on their title and
abstract; of these, 9 (Rydeard et al., 2006; Donzelli et al.,
2006; Anderson, 2005; Gagnon, 2005; Quinn, 2005;
MacIntyre, 2006; Limba da Fonseca et al., 2009; Gladwell
et al., 2006; OBrien et al., 2006) were considered for
inclusion in this review (Fig. 1). Reasons for exclusion are
shown in Fig. 1 for those articles that were excluded in the
screening stage (Curnow et al., 2009; Rodrguez-Fuentes and
Otero-Gargamela, 2009; Ribeiro-Machado, 2006; Ya Li,
2005). Only 1 author did not reply to our inquiries (Curnow
et al., 2009), two authors contacted (Rodrguez-Fuentes
and Otero-Gargamela, 2009; Donzelli et al., 2006) sent
information to us on a new trial for inclusion or exclusion.
A number of trials that were not included in previous
systematic reviews of Pilates-based exercises were
included in this review (Table 1). One new unpublished trial
(MacIntyre, 2006) and one RCT (Limba da Fonseca et al.,
2009) that were not included in any of the previously
published reviews were included in this review (Table 1).
Methodological quality
Agreement among the 2 reviewers was almost high
(ICC Z 0.89) (Atkinson and Nevill, 1998). The methodological quality assessment using the PEDro scale revealed
a mean score of 5.5 (range Z 3e9). One trial (11.1%) did
not consider a point measurement and measurements of
variability. The groups were similar at baseline in 88.8% of
the trials. Blinding of the therapist and blinding of the
subject were not used in any of the trials, as would be
expected for an exercise therapy study. An intention-totreat analysis was used in 22.2% of the trials, allocation
concealment and dropouts were present in 33.3% and 44.4%
of the trials respectively. Eligibility criteria were the only
aspect reported in 100% of the trials. One of the articles
(Donzelli et al., 2006) included in the review received
a score of 3 on the PEDro scale and was the only quasirandomized controlled trial.
Study characteristics
The 9 RCTs included in this review compared Pilates exercise against APT (e.g., therapeutic massage, traditional
dynamic lumbar stabilisation exercises, back school and
standard physiotherapy) or Pilates exercise against MI (e.g.,
no intervention, usual care, normal exercise or sports
regimes) (Table 2). No placebo RCT was identified. Seven
trials assessed the efficacy of Pilates on both variable of
interest (i.e., LBP and FD) (Rydeard et al., 2006; Donzelli
et al., 2006; Anderson, 2005; Gagnon, 2005; MacIntyre,
2006; Gladwell et al., 2006; OBrien et al., 2006). Two
trials just assessed LBP (Limba da Fonseca et al., 2009) or
FD (Quinn, 2005) in that order (Table 1).
Data synthesis and analysis
127
128
Figure 1
Flow chart of systematic review inclusion and exclusion. RCTs Z randomised controlled trials.
Overall the female sample (142 subjects) was twice the

size of the male sample (81 subjects). The age of the
subjects ranged from 30 to 50 y (Table 1).
Pilates regimes were heterogenic in term of frequency
(range Z 1e3 times/wk) and number of sessions
(range Z 6e24). Only in one trial were there 10 consecutive sessions (Donzelli et al., 2006). The typical session
length was usually similar between trials, predominantly
60 min per session, performed on mats (i.e., Pilates matwork) and studio equipment (i.e., Pilates studio apparatus). Rydeard et al. (2006), Donzelli et al. (2006) and
MacIntyre (2006) used specific rehabilitation methods (i.e.,
Biokinetik Exercise Technique, Pilates CovaTech and STOTT
Pilates respectively) derived from the original Pilates
method and its basic principles (Table 2). The use of
certified therapist for Pilates was unreported, except in

one trial (Anderson, 2005).
Four trials used co-interventions (e.g., analgesic
intake, physiotherapy treatment, home exercise programme) in the experimental designs (Rydeard et al.,
2006; Gagnon, 2005; MacIntyre, 2006; Gladwell et al.,
2006). In one trial (Gladwell et al., 2006), Pilates exercise was applied as an additional intervention to the drug
treatment (Table 2).
Pilates versus minimal intervention

Of the 6 trials (169 subjects) included in this treatment
contrast, 5 trials (147 subjects) compared Pilates with MI

Article
Details of the included randomised controlled trials.

Patient characteristics, sample size, and
duration of complaint
Pilates versus another therapeutic treatment

Anderson, 2005
Volunteers recruited from local clinicians
which physicians, surgeons and therapies.
Age 30e58 y
Main exclusion criterion: Significant weakness
of the lower extremities, pregnancy, recent
abdominal surgery.
N Z 21
Gender: 10 Female, 11 Male
Duration of LBP: (mean SD): Pilates group,
18.1 (27.0) mo; massage group, 58 (103.7) mo
Gagnon, 2005
Patients recruited from Sports Medicine Group.
Age 18e45 y
Main exclusion criterion: Positive neural signs,
any serious pathology preventing exercise,
workers compensation or motor vehicle cases.
N Z 12
Duration of LBP: >3 mo in 5 patients, <3 mo in
1 patients for Pilates group >3 mo in 4
patients, <3 mo in 2 for lumbar stabilization
exercises.
Donzelli et al.,
Volunteers recruited from outpatient
2006
departments.
Age 20e65 y
Main exclusion criterion: Structural
deformities, neurological values outside the
normal range, disk hernia, spinal surgery.
N Z 40
Duration of LBP: >3 mo for both groups.a
OBrien et al.,
Patients recruited from medical general
2006
practitioner clinics.
Age 25e65 y
Main exclusion criterion: Nerve root
compression signs, recent spinal fracture,
tumour, or infection; co-morbidities or
contraindications to exercise.
N Z 28
Duration of LBP: overall, 10.9 (7.7) mo.
Gladwell et al.,
Patients recruited via posters, letters given in
2006
clinics and e-mail information.
Age 29e54 y
Main exclusion criterion: Back pain attributed
to any specific pathology. Major surgery within
the past year.
N Z 34
Duration of LBP: Pilates group, 115 (101) mo;
minimal intervention group, 139 (148) mo.
Outcomes (measure) Article included in

previous reviews
Pain (MBI-Pain)
Disability (ODQ)
Included in Lim et al.,

2011
Pain (VAS)
Disability (ODQ)

2011
Pain (VAS)
Disability (ODQ)
Included in La Touche
et al., 2008 and Lim
et al., 2011
Pain (VAS)
Disability (RMDQ)
Included in Posadzki
et al., 2011
Pain (VAS)
Disability (ODQ)
et al., 2008; Posadzki
et al., 2011
(continued on next page)
Table 1
129

130
Table 1 (continued )
Article
Patient characteristics, sample size, and

duration of complaint
Outcomes (measure) Article included in

previous reviews
Limba da
Fonseca
et al., 2009
Patients recruited from waiting list for

physiotherapeutic treatment
Age 21e47 y
Main exclusion criterion: prior back surgery or
radiological signs of spinal instability.
N Z 14
Duration of LBP 6 moa
Subjects were recruited through notices posted
to private and public physicians and
physiotherapists offices, local sports clubs and
Universities.
Age 20e55 y
Main exclusion criterion: prior spinal surgery or
spinal fracture, evidence of overt neurological
compromise or acute inflammatory process,
pregnant.
N Z 39
Duration of LBP: (mean [range]): Pilates group,
66 [6e324] mo; minimal intervention, 108 [12
e240] mo.
Participant recruited voluntarily from a private
physiotherapy practice.
Age 25e62 y
Main exclusion criterion: Previous or current
participation in a Pilates or back class program,
spinal surgery
N Z 32
Duration of LBP: Pilates group, 27.0 (22.2) mo;
minimal intervention, 20.4 (19.0) mo
Volunteers recruited from local commercial
and community-fitness center via flyers.
Age 27e53
Main exclusion criterion: No specified
N Z 22
Gender: No specified
Duration of LBP 6 moa
Patients recruited from medical general
practitioner (GP) clinics.
Age 25e65 y
Main exclusion criterion: Nerve root
compression signs, recent spinal fracture,
tumour, or infection; co-morbidities or
contraindications to exercise.
N Z 28
Duration of LBP: overall, 10.9 (7.7) mo.
Pain (VAS)
Rydeard et al.,
2006
MacIntyre,
2006
Quinn, 2005
OBrien et al.,
2006
Not included in previous

reviews
Pain (NRS-101)
Disability (RMDQ-HK) et al., 2008; Posadzki
et al., 2011
Pain (VAS)
Disability (RMDQ)
Not included in previous

reviews
Disability (ODQ)

2011
Pain (VAS)
Disability (RMDQ)
Included in Posadzki
et al., 2011
Abbreviations: VAS Z visual analog scale, MBI-Pain: Miami Back Index Pain Sub-Scale, NRS-101: 101-point numerical rating scale,
ODQ Z Oswestry Disability Questionnaire, RM Z Roland-Morris Disability Questionnaire, RM-HK Z RolandeMorris Questionnaire Chinese
version.
a
Duration of complaint as inclusion criterion.

Article
Details of the interventions and main outcomes in the included randomised controlled trials.
Interventions

Anderson 2005
Experimental group: Pilates
apparatus.
Regime: 50 min., 2 time/
wk, 12 sessions.
Comparison group:
therapeutic massage.
Regime: 30 min, 2 time/wk,
12 sessions.
Gagnon 2005
matwork.
Regime: 30e45 min., 1.43
time/wk, 10.5 sessions and
PT.
Comparison group:
traditional dynamic lumbar
stabilization exercises.
Regime: 30e45 min., 1.46
time/wk, 9.67 sessions and
PT.
Donzelli et al.,
Experimental group:
b
2006
Pilates CovaTech matwork.
Regime: 60 min., 10.5
consecutive sessions.
Comparison group: Back
School.
Regime: 60 min., 10.5
consecutive sessions.
OBrien et al.,
2006
matwork and apparatus.
wk, 8 sessions
Comparison group: standard
physiotherapy (i.e., manual
therapy, education, core
stability exercises,
stretches, McKenzie,
interferential, orthotics,
taping and/or laser).
wk, 8 sessions
Gladwell et al.,
2006
matwork as an additional
intervention.
Regime: 60 min, 1 times/
wk, 6 sessions.
Comparison group: normal
daily activities, drug (i.e.,
analgesics) and no exercise.
Limba da
Fonseca
matwork
et al., 2009
Regime: 60 min. 2 times/
wk, 15 sessions.
Comparison group:
maintenance normal daily
activities, non-drugs.
Preintervention
Postintervention
PEDro scorec
Pilates: MBI-pain, 33.5

(18.6); ODQ, 16.7 (4.2)
Therapeutic massage: MBIpain, 39.3 (15.6); ODQ, 18.5
(5.9)
Pilates: MBI-pain, 24.2

(14.7); ODQ, 13.9 (5.7)
Therapeutic massage: MBIpain, 35.0 (18.0); ODQ, 17.9
(7.2)
Pilates: VAS, 2.0 (1.7); ODQ,

15.8 (3.7)
Lumbar stabilization
exercise: VAS, 3.8 (2.5);
ODQ, 17.2 (6.1)
Pilates: VAS, 0.9 (1.7); ODI,

7.0 (5.9)
Lumbar stabilization
exercise: VAS, 1.5 (1.7);
ODQ, 9.1 (7.5)
Pilates CovaTech: VAS, 7.3

(2.2); ODQ, 13.6 (7.0)
Back School: VAS, 6.8 (3.1);
ODQ, 10.0 (6.5)
Pilates CovaTech: VAS, 4.5

(2.1); ODQ, 6.9 (3.9)
Back School: VAS, 4.3 (3.0);
ODQ, 7.7 (6.2)
Pilates: VAS, 61.6 (6.5);

RMDQ, 10.7 (1.9)
Standard physiotherapy:
VAS, 54.7 (6.5); RMDQ, 9.5
(1.8)

RMDQ, 1.7 (0.7)
Standard physiotherapy:
VAS, 17.8 (6.9); RMDQ, 4.2
(1.5)

19.7 (9.8)
Control: VAS, 2.4 (0.9);
ODQ, 24.1 (13.4)

18.1 (11.2)
ODQ, 18.1 (13.0)
Pilates: VAS, 5.9 (2.0)

Control: VAS, 6.1 (1.8)
Pilates: VAS, 3.0 (3.4)

Control: VAS, 4.9 (2.5)
(continued on next page)
Table 2
131

132
Table 2 (continued )
Article
Interventions
Preintervention
Postintervention
PEDro scorec
Rydeard et al.,
2006

wk, 12 sessions, plus home
exercise program.
Regime: 15 min. 6 times/wk
Comparison group: health
usual care, analgesics, PT as
needed and no exercise.
Experimental group: aSTOTT
Pilates matwork.
Regime: 60 min.1 time/wk,
12 sessions and normal
exercise or sports regimes,
PT as needed; plus home
exercise program
wk, 10 sessions.
Comparison group: normal
exercise or sports regimes
and PT as needed.
matwork.
Regime: 45e60 min., 2
time/wk, 24 sessions.
Comparison group:
maintenance normal daily
activities and no exercise.
wk, 8 sessions
Comparison group: no
exercise.
Pilates: NRS-101, 23.0 (1.7);

RMDQ, 3.1 (2.7)
Control: NRS-101, 30.4
(1.7); RMDQ, 4.2 (3.3)
Pilates: NRS-101, 18.3 (1.4);

RMDQ, 2.0 (1.3)
Control: NRS-101, 33.9
(1.4); RMDQ, 3.2 (1.7)

RMDQ, 7.0 (3.1)
RMDQ, 7.4 (3.4)

RMDQ, 4.6 (2.3)
RMDQ, 2.2 (3.6)
Pilates: ODQ, 25.9 (10.7)

Control: ODQ, 22.0 (8.7)
Pilates: ODQ, 10.9(10.3)

Control: ODQ, 18.0 (12.4)
Pilates: VAS, 61.6(6.5);

RMDQ, 10.7 (1.9)
RMDQ, 7.0 (1.9)
Pilates: VAS, 17.2(7.7);

RMDQ, 1.7 (0.7)
RMDQ, 6.2 (1.5)
MacIntyre,
2006
Quinn, 2005
OBrien et al.,
2006
Abbreviation: PT Z Physiotherapeutic treatments.

a
STOTT Pilates is an anatomically-based approach to the original method that emphasizes neutral alignment, core or trunk stability
and peripheral mobility.
b
Pilates CovaTech is a specific rehabilitation method derived from the original Pilates method in Italy and takes the name from the
therapist who invented it.
c
Higher scores indicating better methodological qualities of trials.
for pain relief (Rydeard et al., 2006; MacIntyre, 2006;

Limba da Fonseca et al., 2009; Gladwell et al., 2006;
OBrien et al., 2006) and 5 trials (155 subjects) for
reducing disability (Rydeard et al., 2006; Quinn, 2005;
MacIntyre, 2006; Gladwell et al., 2006; OBrien et al.,
2006). The methodological quality of the articles ranged
from 5 to 8 (Table 2). The pooled results of comparing
Pilates exercise with MI showed a statistically significant
difference for pain relief (pooled ES weighted Z 0.44,
95% CI Z 0.09 to 0.80) (Fig. 2). This difference in pain
score was moderate according to ES interpretation (i.e.,
ES 0.2 poor or weak; between 0.3 and 0.7 moderate; and
0.8 high or strong) (Thomas et al., 2005). When estimating the publication bias for this treatment contrast, 9
studies with EStrivial Z 0.15 would have to be present, for

the results to be threatened.
There was no evidence that Pilates exercise was effective for improving disability (pooled ES weighted Z 0.28,
95% CI Z 0.07 to 0.62) (Fig. 2).
Pilates versus another physiotherapeutic treatment

Of the 9 studies included in this review, 4 trials (101
patients) compared Pilates with APT in both variables of
interest (Donzelli et al., 2006; Anderson, 2005; Gagnon,
2005; OBrien et al., 2006). The methodological quality
of the articles ranged from 3 to 7 (Table 2). There was

133
Figure 2 Forest plot of the results of trials comparing Pilates-based exercises with minimal interventions for LBP and FD. Values
presented are effect size (with correction factor) and 95% confidence interval. The pooled effect sizes were calculated using
a weighting by variance reciprocal.
a moderate and significant difference between treatment

groups
in
disability
score
(pooled
ES
weighted Z 0.55, 95% CI Z 0.08 to 1.03) (Fig. 3). For
this treatment contrast, 10 studies would have to be
unpublished with EStrivial Z 0.15, for the outcome to be
threatened.
The pooled results of the comparison between Pilates
exercise and APT did not show statistically significant
differences for pain relief (pooled ES weighted Z 0.14,
95% CI Z 0.27 to 0.56).
Results for testing homogeneity and weighted

regression
The test for homogeneity showed that treatment contrast
for pain relief: Pilates against MI [H Z 125.29 >
2
0.95X Z 9.48 with 4 degrees or freedom] and treatment
contrast
for
disability:
Pilates
against
ATT
[H Z 101.95 > 0.95X2 Z 7.81 with 3 degrees of freedom] did
not represent a similar measure of treatment effectiveness.
When the effects of particular features prior coding
(e.g., quality of the trials, duration of complaint, regime of
Pilates, performed on mats as opposed to studio equipment, age of the patients, BMI, use of certified therapist)
on treatment contrast that were statistically significant,
were evaluated, the statistics model used did not detect
any predictor variable (Table 3).
Discussion
This systematic review provides evidence that Pilates-based
exercise is moderately superior to minimal intervention for
pain relief in subjects with persistent, non-specific LBP
(Fig. 2). However, the presence of co-interventions (e.g.,
analgesic intake, physiotherapy treatment as needed) in the
Pilates exercise groups of 2 trials (MacIntyre, 2006; Gladwell
et al., 2006) was an aspect that could have raised the effect
magnitude, contributing to the increased pain relief. This
methodological issue compromised internal validity; therefore, when assessing the impact of the co-interventions on
this treatment contrast, a sensitivity analysis with exclusion
of trials with co-interventions (MacIntyre, 2006; Gladwell
et al., 2006) was performed. When the trials with cointerventions were deleted, the effect size unexpectedly
increased considerably (i.e., pooled ES weighted Z 1.12,
95% CI Z 0.60 to 1.64) for pain relief. Nevertheless, this
was the finding of 3 trials, and more research is needed to
confirm the results.
The pooled results showed Pilates-based exercise is
moderately superior to another physiotherapy treatment in
reducing disability. Interestingly, this finding was based on
pooled data from 4 different trials (i.e., therapeutic massage,
traditional dynamic lumbar stabilisation exercises, back
school and standard physiotherapy) where Pilates was not
superior to these particularly prescribed treatments (Table 2).
On the other hand, these pooled results were biased by the
134
Figure 3 Forest plot of the results of trials comparing Pilates-based exercises with another physiotherapeutic treatment for LBP
and FD. Values presented are effect size (with correction factor) and 95% confidence interval. The pooled effect sizes were
calculated using a weighting by variance reciprocal.
used of co-intervention in 1 trial (Gagnon, 2005), but, when

this study was deleted, the effect size unexpectedly
increased slightly (i.e., pooled ES weighted Z 0.62, 95%
CI Z 0.10 to 1.14) for the disability score.
This systematic review did not find evidence that
Pilates was superior to minimal intervention in reducing
disability, nor superior to another physiotherapy
treatment for pain relief, because the pooled results did
not show statistically significant differences (Figs. 2
and 3).
When looking at the quality of the trials included in this
review, a mean score of 6 can be considered a moderate
score because these trials were exercise trials where it is
impossible to blind the treatment provider and subjects,
and, therefore, the maximum PEDro score that can be
achieved is 8. Trials that associated Pilates against minimal
intervention scored >6. However, trials associating Pilates
against another physiotherapy treatment were of lower
methodological quality (mean score Z 5), and according to
CBRG (Furlan et al., 2009) potentially presented biased
estimates of treatment effectiveness. The implications
associated with the poor methodological quality of the
studies reviewed affecting the observed effects were welldescribed in previous systematic reviews (Lim et al., 2011;
Pereira et al., 2011).
On the other hand, the test for homogeneity showed

there was some source of additional variation that could
be related to the effect of moderator variables. An
intriguing finding was that the effect sizes did not deviate
substantially from the weighted regression model. This
outcome suggests that the small number of trials
included in the study limits the potential merits of
a meta-regression approach. It is also possible that
several features not coded a priori that could have
influenced the treatment effectiveness, were linked to
psychological impairments (e.g., pain perception, fearavoidance behaviour, neurological feedback from the
deep muscles of the trunk, low self-efficacy and catastrophizing), physical impairments (e.g., weakness,
endurance and structural integrity), characteristics of
treatment implementation (e.g., the experience of the
Pilates instructor in the management of the special
characteristic of the subjects, progression), pain coping
strategies (e.g., behavioural and cognitive approach,
behavioural and cognitive avoidance), environmental and
social factors.
The main limitations of this meta-analysis were in the
review process (e.g., only a few studies assessing the most
important outcomes of interest) and generalizability (e.g.,
limited data for persistent LBP and the review was not able

Table 3
135
Summary of the regression model for sizes effect significantly different of zero.
Particular variables
Pilates versus minimal interventions

R2
F-value
P-value
Age
BMI
Duration of complaint
Quality of the study
Number of session of Pilates
Frequency of the session (time/wk)
Duration of the session (in min)B
Type of Pilates class
Certified instructors
0.033
0.046
0.074
0.044
0.003
0.210
e
0.310
0.757
0.102
0.097
0.240
0.137
0.008
0.797
e
1.347
9.337
0.771
0.785
0.657
0.736
0.935
0.438
e
0.330
0.055
5
4
5
5
5
5
e
5
5
Particular variables

Disability
Age
BMI5
Duration of complaint
Quality of the study
Number of session of Pilates
Frequency of the session (time/wk)
Duration of the session (in min)B
Type of Pilates class
Certified instructors
R2
F-value
P-value
0.0393
e
0.393
0.702
0.011
0.363
0.051
0.007
0.586
1.294
e
1.315
4.713
0.022
1.139
0.108
0.014
2.836
0.373
e
0.370
0.162
0.896
0.398
0.773
0.915
0.234
4
e
4
4
4
4
4
4
4
Abbreviations: k Z number of trials; BMI Z Body mass index;

reviewed.
Constant variable for the trials reviewed;
to offer dosage parameters). The authors did not limit the

set for the quality assessment score of the RCTs, because it
would restrict the possibility of accomplishing metaregression. Limitations between studies were the small
sample sizes, the large variation in intervention protocols
(e.g., Pilates regimes: frequency and progression, and
Pilates type: performed on mats as opposed to studio
equipment), the absence of placebo controlled trials, the
short-term follow-up and the absence of assessment of
treatment adherence.
When comparing the strengths of this review in relation
to other studies, the authors limited the set for study
design (e.g., RCT), and detected that there is a source of
additional variation that can affect the independent
manner the effectiveness of the treatment, and measured
how many studies were involved in the file drawers
problem. The results found in this systematic review were
similar to those presented by Lim et al. (2011). In addition,
this review included 2 new trials (MacIntyre, 2006; Limba
da Fonseca et al., 2009) that were not included by Lim
et al. (2011), accounting for the addition of 49 subjects.
However, the results found in this systematic review were
different to those presented by Pereira et al. (2011).
Basically, the difference was determined by the inclusion
in this review, of studies with serious methodological
flaws.
In the authors opinion, future trials could also consider
the preventive actions of Pilates-based exercise to reduce
episodes of LBP and associated functional disability.
Not reported on the trials
Conclusions
The results of this systematic review suggest that Pilatesbased therapeutic exercise is moderately superior to minimal
intervention for pain relief and confers similar benefits when
compared with pooled scores to another physiotherapeutic
treatment. Although co-interventions with Pilates, might
enhance treatment effectiveness for pain relief; this
conclusion should be interpreted with caution.
Pilates is moderately better than another physiotherapeutic treatment in reducing disability, and provides
comparable benefits to minimal intervention. The low
methodological quality of the studies reviewed, and the
heterogeneity of the physiotherapy treatment showed
estimate bias of the treatment effectiveness for reducing
disability; thus, it is recommended that Pilates should be
carefully considered for patients with functional disability
associated to LBP.
The optimal implementation of Pilates exercise at
present is unclear. Future studies should incorporate
placebo controlled trial, larger sample sizes, intervention
protocols that are comparable, assessment of the several
features not coded in this review and longer term follow-up.
Conflict of interest
None.
Pain

136
Appendix A. Supplementary material
Supplementary material associated with this article can be

found, in the online version, at http://dx.doi.org/10.1016/
j.jbmt.2012.08.003.
References
Akuthota, V., Nadler, S.F., 2004. Core strengthening. Arch. Phys.
Med. Rehabil. 85 (3 Suppl. 1), S86eS92.
Anderson, G.B.J., 1999. Epidemiological features of chronic lowback pain. Lancet 354, 581e585.
Anderson, B.D., 2005. Randomized clinical trial comparing active
versus passive approaches to the treatment of recurrent and
chronic low back pain. Dissertation. University of Miami, Miami,
FL, p. 225.
Anderson, B., Spector, A., 2000. Introduction to Pilates-based
rehabilitation. OCL 9, 395e410.
Asghari, A., Nicholas, M., 2001. Pain self-efficacy beliefs and pain
behaviour. A prospective study. Pain 94 (1), 85e100.
Atkinson, G., Nevill, A.M., 1998. Statistical methods for assessing
measurement error (reliability) in variables relevant to sports
medicine. Sports Med. 26 (4), 217e238.
Bombardier, C., van Tulder, M.W., Pennick, V., et al., 2006.
Cochrane Back Group. About the Cochrane Collaboration.
Cochrane Review Groups [CRGs], p. 4.
Bryan, M., Hawson, S., 2003. The benefits of Pilates exercise in
orthopaedic rehabilitation. Tech. Orthop. 18 (1), 126e129.
Curnow, D., Cobbin, D., Wyndham, J., et al., 2009. Altered motor
control, posture and the Pilates method of exercise prescription. J. Bodywork Move. Ther. 13, 104e111.
Donzelli, S., Di Domenica, F., Cova, A.M., et al., 2006. Two
different techniques in the rehabilitation treatment of low back
pain: a randomized controlled trial. Eura. Med. 42, 205e210.
Furlan, A.D., Pennick, V., Bombardier, C., van Tulder, M., 2009.
Updated method guidelines for systematic reviews in the
Cochrane back review group. Spine 34, 1929e1941.
Gagnon, L.H., 2005. Efficacy of Pilates exercises as therapeutic
intervention in treating patients with low back pain. Dissertation. University of Tennessee, Knoxville, TN, p. 119.
Gladwell, V., Head, S., Haggar, M., et al., 2006. Does a program of
Pilates improve chronic non-specific low back pain? J. Sport
Rehabil. 15, 338e350.
Hart, L.G., Deyo, R.A., Cherkin, D.C., 1995. Physician office visits
for low back pain frequency, clinical evaluation, and treatment
patterns from a US National Survey. Spine 20 (1), 11e19.
Hedges, L.V., Olkin, I., 1985. Statistical Methods for Meta-analysis.
Academic Press, New York.
Hedges, L.V., 1981. Distribution theory for Glasss estimator of
effect size and related estimator. J. Educ. Statistic 6, 107e128.
Kirkaldy-Willis, W., Bernard, T.N., 1999. A comprehensive outline
of treatment. In: Kirkaldy-Willis, W., Bernard, T.N. (Eds.),
Managing Low Back Pain. Churchill Livingstone, Philadelphia.
La Touche, R., Escalante, K., Linares, M.T., 2008. Treating nonspecific chronic low back pain through the Pilates method. J.
Bodywork Move. Ther. 12, 364e370.
Lackner, J.M., Carosella, A.M., Feuerstein, M., 1996. Pain expectancies, pain, and functional self-efficacy expectancies as
determinants of disability in patients with chronic low back
disorders. J. Consult. Clin. Psychol. 64 (1), 210e220.
Lim, E.C., Poh, R.L., Low, A.Y., Wong, W.P., 2011. Effects of
Pilates-based exercises on pain and disability in persistent
nonspecific low back pain: a systematic review with metaanalysis. J. Orthop. Sports Phys. Ther. 41, 70e80.
Limba da Fonseca, J., Magini, M., Helena de Freitas, T., 2009.
Laboratory gait analysis in patients with low back pain before
and after a Pilates intervention. J. Sport Rehabil. 18, 269e282.

Luo, X., Pietrobon, R., Sun, S.X., et al., 2004. Estimates and
patterns of direct health care expenditures among individuals
with back pain in the United States. Spine 29 (1), 79e86.
MacIntyre, L., 2006. The effect of Pilates on patients chronic low
back pain. A pilot study. Dissertation. University of the Witwatersrand, Johannesburg, p. 59.
Maher, C.G., Sherrington, C., Herbert, R.D., et al., 2003. Reliability
of the PEDro scale for rating quality of randomized controlled
trials. Phys. Ther. 83, 713e721.
Maher, C.G., 2004. Effective physical treatment for chronic low
back pain. Orthop. Clin. North Am. 35, 57e64.
Mannion, A.F., Junge, A., Taimela, S., et al., 2001. Active therapy
for chronic low back pain: part 3. Factors influencing self-rate
disability and its change following therapy. Spine 26 (8),
920e929.
Moher, D., Liberati, A., Tetzlaff, J., Altman, D.G.PRISMA Group,
2009. Preferred reporting items for systematic reviews and
meta-analyses: the PRISMA statement. BMJ 339, b2535.
Muscolino, J.E., Cipriani, S., 2004. Pilates and the powerhouse.
J. Bodywork Move. Ther. 8 (1), 15e24.
OBrien, N., Hanlon, M., Meldrum, D., 2006. Randomized controlled
trial comparing physiotherapy and Pilates in the treatment of
ordinary low back pain. Phys. Ther. Rev. 11, 224e225.
Pereira, L.M., Obara, K., Dias, J.M., Menacho, M.O.,
Guariglia, D.A., Schiavoni, D., Pereira, H.M., RosaCardoso, J., 2011. Comparing the Pilates method
with no exercise or lumbar stabilization for pain and functionality in patients with chronic low back pain:
systematic review and meta-analysis. Clin. Rehabil., 1e11.
http://dx.doi.org/10.1177/0269215511411113.
Posadzki, P., Lizis, P., Hagner-Derengowska, M., 2011. Pilates for
low back pain: a systematic review. Complement. Ther. Clin.
Pract. 17 (2), 85e89.
Quinn, J.V., 2005. Influence of Pilates-based mat exercise on
chronic lower back pain. Dissertation. Florida Atlantic University, Boca Raton, FL, p. 74.
Ribeiro-Machado, C.A.N., 2006. Effect of one program established
in the Pilates method in the reduction of the lombalgia in the
gestation. Fisioterapia Brasil 7 (5), 345e350.
Rodrguez-Fuentes, G., Otero-Gargamela, M.D., 2009. Selfperception of backaches improvement on patients who practice
Pilates method. Personal communication with the authors.
Rydeard, R., Leger, A., Smith, D., 2006. Pilates-based therapeutic
exercise: effect on subjects with nonspecific chronic low back
pain and functional disability: a randomized controlled trial.
Orthop. Sports Phys. Ther. 36 (7), 472e484.
Sherrington, C., Herbert, R.D., Maher, C., et al., 2000. PEDro:
a database of randomized controlled trials and systematic
reviews in physiotherapy. Man. Ther. 5, 223e226.
Sorosky, S., Stilp, S., Akuthota, V., 2008. Yoga and Pilates in the
management of low back pain. Curr. Rev. Musculoskelet. Med.
1, 39e47.
Thomas, J.R., French, K.E., 1986. The use of meta-analysis in
exercise and sport: a tutorial. Res. Q. Exerc. Sport 57 (3),
196e204.
Thomas, J.R., Nelson, J.K., Silverman, S.J., 2005. Research Methods
in Physical Activity, fifth ed. Human Kinetics, Champaign.
Waddell, G., Newton, M., Henderson, I., et al., 1993. A fearavoidance beliefs questionnaire (FABQ) and the role of fearavoidance beliefs in chronic low back pain and disability. Pain
52 (2), 157e168.
Woby, S., Watson, T., Roach, N., et al., 2004. Are changes in fear
avoidance beliefs, catastrophizing and appraisals of control,
predictive of changes in chronic low back pain and disability?
Eur. J. Pain 8, 201e210.
Ya Li, T., 2005. The effects of Pilates training on patients with low
back pain. Dissertation. National University of the Education of
Pingtung. Taiwan, p. 88.

Aladro Et Al., (2012) - Pilates-Based Exercise For Persistent, Non-Specific Low Back Pain and Associated Functional Disability. A Meta-Analysis With Meta-Regression

Загружено:

Сведения о документе

Оригинальное название

Авторское право

Доступные форматы

Поделиться этим документом

Поделиться или встроить документ

Параметры публикации

Этот документ был вам полезен?

Это неприемлемый материал?

Авторское право:

Доступные форматы

Aladro Et Al., (2012) - Pilates-Based Exercise For Persistent, Non-Specific Low Back Pain and Associated Functional Disability. A Meta-Analysis With Meta-Regression

Загружено:

Авторское право:

Доступные форматы

This article appeared in a journal published by Elsevier.

Author's personal copy

Available online at www.sciencedirect.com

PREVENTION & REHABILITATION e PILATES EXERCISE: META ANALYSIS

Pilates-based exercise for persistent, non-specific

about its frequency, recurrence, treatment and cost

PREVENTION & REHABILITATION e PILATES EXERCISE: META ANALYSIS

journal homepage: www.elsevier.com/jbmt

Author's personal copy

PREVENTION & REHABILITATION e PILATES EXERCISE: META ANALYSIS

A.R. Aladro-Gonzalvo et al.

Data extraction and quality assessment

Author's personal copy

The data were grouped into two treatment contrasts: (1)

meta-analysis, and dES trivial is the estimate of a trivial ES

PREVENTION & REHABILITATION e PILATES EXERCISE: META ANALYSIS

Data synthesis and analysis

Author's personal copy

PREVENTION & REHABILITATION e PILATES EXERCISE: META ANALYSIS

A.R. Aladro-Gonzalvo et al.

Overall the female sample (142 subjects) was twice the

certified therapist for Pilates was unreported, except in

Pilates versus minimal intervention

Author's personal copy

Details of the included randomised controlled trials.

Pilates versus another therapeutic treatment

Outcomes (measure) Article included in

Included in Lim et al.,

Included in Lim et al.,

(continued on next page)

PREVENTION & REHABILITATION e PILATES EXERCISE: META ANALYSIS

Author's personal copy

A.R. Aladro-Gonzalvo et al.

PREVENTION & REHABILITATION e PILATES EXERCISE: META ANALYSIS

Patient characteristics, sample size, and

Outcomes (measure) Article included in

Patients recruited from waiting list for

Not included in previous

Not included in previous

Included in Lim et al.,

Author's personal copy

Pilates versus another therapeutic treatment

Pilates: MBI-pain, 33.5

Pilates: MBI-pain, 24.2

Pilates: VAS, 2.0 (1.7); ODQ,

Pilates: VAS, 0.9 (1.7); ODI,

Pilates CovaTech: VAS, 7.3

Pilates CovaTech: VAS, 4.5

Pilates: VAS, 61.6 (6.5);

Pilates: VAS, 17.2 (7.7);

Pilates: VAS, 2.7 (0.9); ODQ,

Pilates: VAS, 2.2 (0.9); ODQ,

Pilates: VAS, 5.9 (2.0)

Pilates: VAS, 3.0 (3.4)

(continued on next page)

PREVENTION & REHABILITATION e PILATES EXERCISE: META ANALYSIS

Author's personal copy

A.R. Aladro-Gonzalvo et al.

PREVENTION & REHABILITATION e PILATES EXERCISE: META ANALYSIS

Experimental group: Pilates

Pilates: NRS-101, 23.0 (1.7);

Pilates: NRS-101, 18.3 (1.4);

Pilates: VAS, 5.1 (2.0);