Accepted Manuscript

Centralization and directional preference: An updated systematic review with

synthesis of previous evidence

Stephen May, Nils Runge, Alessandro Aina

PII: S2468-7812(18)30206-6
DOI: 10.1016/j.msksp.2018.09.006
Reference: MSKSP 1930

To appear in: Musculoskeletal Science and Practice

Received Date: 1 June 2018

Revised Date: 29 August 2018
Accepted Date: 6 September 2018

Please cite this article as: May, S., Runge, N., Aina, A., Centralization and directional preference: An
updated systematic review with synthesis of previous evidence, Musculoskeletal Science and Practice
(2018), doi: https://doi.org/10.1016/j.msksp.2018.09.006.

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.
 ACCEPTED MANUSCRIPT
Title page

Centralization and directional preference: an updated systematic review with synthesis of

previous evidence

PT
RI
Stephen May1*, Nils Runge1, Alessandro Aina2

SC
s.may@shu.ac.uk

U
Nils.A.Runge@student.shu.ac.uk AN
sandroaina@gmail.com
M

1
Faculty of Health and Wellbeing, Sheffield Hallam University, Sheffield, UK
2
Studio di Riabilitazione MDTC, Milano, Italy
D
TE

1
* corresponding author

Faculty of Health and Wellbeing, Sheffield Hallam University, Sheffield, S10 2 BP, UK
EP

+44 0114 225 2370

C

s.may@shu.ac.uk
AC
 ACCEPTED MANUSCRIPT

Background: Centralization and directional preference are common

management and prognostic factors in spinal symptoms.

Objective: To update the previous systematic review.

Design: Systematic review to synthesis multiple aspects of

PT
centralization and directional preference.

Method: Contemporary search was made of multiple databases using

RI
relevant search terms. Abstracts and titles were filtered by two authors;
relevant articles were independently reviewed by two authors for content,

SC
data extraction, and quality.

Results: Forty-three additional relevant articles were found. The

U
quality of the studies, using PEDro for randomized controlled trials, was
AN
moderate or high in six out of ten RCTs; moderate or high in six out of 12
cohort studies. Prevalence of centralization was 40%, the same as the previous
review. Directional preference prevalence was only 26%, much lower than the
M

previous review; but neither clinical response was recorded in about a third of
patients. Centralization and directional preference were confirmed as key
D

positive prognostic factors, certainly in patients with low back pain, but limited
TE

evidence for patients with neck pain. There was no evidence that these might
be important treatment effect modifiers. One study evaluated reliability, and
found generally poor levels, despite training.
EP

Conclusions: Centralization and directional preference are worthwhile

indicators of prognosis, and should be routinely examined for even in patients
C

with chronic low back pain. But they do not occur in all patients with spinal
AC

problems, and there was no evidence that they were treatment effect
modifiers.
 ACCEPTED MANUSCRIPT
Centralization and directional preference: An updated systematic review

with synthesis of previous evidence

PT
1. Introduction

RI
Centralization is defined as the abolition of distal and spinal pain in response to

repeated movements or sustained postures (McKenzie and May, 2003).

SC
Directional preference is defined as the repeated movement that produces

U
centralization, or an abolition or decrease in symptoms, or an increase in
AN
restricted range of movement (McKenzie and May, 2003). Centralization and
M

directional preference are thus important clinical phenomena, as they occur in

D

response to therapeutic loading strategies and thus are clinically induced, and
TE

also as they describe a lasting change. They have been commonly referenced

(May and Aina, 2012), in fact centralization is probably the most commonly
EP

spoken of clinically induced symptom response in the literature. Furthermore

C

they are potentially useful prognostic and management indicators (May and
AC

Aina, 2012). For instance, the presence of centralization has been associated

with better pain, function, return to work, and non-surgical outcomes both

short and long-term (Long, 1995; Werneke and Hart, 2001; Skytte et al., 2005).

Whilst directional preference has been a useful indicator of appropriate

exercises, compared to other treatments in the short-term (Delitto et al., 1993;

 ACCEPTED MANUSCRIPT
Long et al., 2004, 2009). There is the suggestion that centralization and

directional preference maybe helpful in determining the most effective

management strategy (Long et al., 2004); thus being a so-called treatment

effect modifier (May and Aina, 2012).

PT
Centralization has been the subject of several systematic reviews (Aina et al.,

RI
2004; Chorti et al., 2009; May and Aina, 2012). All these are reasonably dated;

SC
even the most recent is over five years old now (May and Aina, 2012).

Although now somewhat dated that last review included 54 studies relevant to

U
AN
centralization and eight studies relevant to directional preference. Since the

inclusion date of that last review a number of additional studies have been
M

published (for instance, Albert et al., 2011; Petersen et al., 2011; Edmond et
D

al., 2014; Apeldoorn et al., 2016). Furthermore, a recent systematic review of

TE

49 articles of relevant prognostic factors, concluded that there was

EP

inconsistent evidence for the usefulness of most clinical findings, centralization

and non-organic signs being the exceptions (Hartvigsen et al. 2015). Thus there
C
AC

are very limited tools for the clinician to determine if a patient might recover

or not, and thus is an important area for further study.

In terms of updating systematic reviews, it has been suggested that the median

survival time of a systematic review is 5.5 years with 23% being outdated

within two years of publication (Shojania et al., 2007). A recent consensus

 ACCEPTED MANUSCRIPT
statement and check list focussed on the question of whether, when and how

to update a previous systematic review (Garner et al., 2016). It did not

stipulate how the results from the previous review(s) should be included;

whether to summarise the conclusions of the previous review only, or whether

PT
to amalgamate the earlier individual studies with the new ones.

RI
Thus it would appear to be appropriate to update the previous systematic

SC
reviews. The aims of the present paper were to summarise previous findings

and to systematically review recent literature since June 2011 relating to all

U
AN
aspects of centralization and directional preference.

2. Methods
M
D

2.1 Study selection and reporting

TE

Any full-text study that reported some aspect of centralization or directional

EP

preference, in adults reporting low back or neck pain, with or without radiating

symptoms was included. PRISMA guidelines for reporting systematic reviews

C

were followed (Harms, 2009). For data prior to the present search (June 2011
AC

to December 2017) data were synthesised from the previous review (May and

Aina, 2012).

2.2 Data sources and searches

 ACCEPTED MANUSCRIPT
A search was made of Medline, Cinahl and AMed from June 2011, the date of

the last search, until December 2017. The website www.mckenziemdt.org,

which lists references relevant to the McKenzie method, and includes a section

on centralization, was also used. The reference lists of all included articles

PT
were also searched. Search terms used were as follows: centralization, OR

RI
directional preference; OR phenomenon; AND spine pain, OR back pain, OR

SC
neck pain, OR cervical, OR lumbar; used individually and in combination. Titles

and abstract were reviewed initially by one author (AA) to see if they might be

U
relevant and duplications removed. All potential articles were reviewed by two
AN
authors (AA, SM) to determine their final relevance, with any disagreements
M

resolved by referral to the third author.

D

2.3 Data extraction and quality assessment

TE

Data extraction was done independently and blinded to each other by two
EP

authors (SM, NR); any disagreements were resolved by consensus. Likewise

C

quality assessment was done independently and blinded to each other by two
AC

authors (SM, NR); any disagreements were resolved by consensus. In both

instances there was the option, not used, to refer to the third author if a

consensus was not reached.

There is not full agreement on the best methods to assess quality either in

randomized controlled trials (RCTs), or in cohort studies. There are many tools
 ACCEPTED MANUSCRIPT
to choose from for such tasks (Sanderson et al., 2007; Olivio et al., 2008), but

the different criteria share may commonalities. The Physiotherapy Evidence

Database (PEDro) scale was used to assess internal and external validity of the

RCTs. The PEDro scale is comprised of 11 criteria (only 10 of which are scored),

PT
has been shown to be valid and reliable, and was used in a recent systematic

RI
review, with a score of 7 or above considered high, 5 or 6 moderate, and 4 or

SC
below poor quality (Young et al., 2018). To assess the quality of cohort and

observation studies a tool for prognostic studies was used that had been

U
adopted from earlier work by Hayden et al. (2013), and was used in a recent
AN
systematic review (Hartvigsen et al., 2015). The quality criteria consist of five
M

domains, with 15 items, which was scored as yes, OR no / unsure / not stated,
D

so that we had a dichotomous outcome; and led to an overall scoring of low,

TE

moderate or high risk of bias; this decision being reached by consensus.

Besides the dichotomous outcome, other minor amendments were made to

EP

the criteria for clarification: criterion 3: 'population of interest' was changed to

C

'source population'; criteria 4 and 5: completeness of follow-up was defined as

AC

one-year and 85% of the inception cohort; criterion 14: we made to include the

ability 'to account for other prognostic factors', as in a multivariate analysis.

Assessing methodological quality in the other studies was not possible due to

the range of study designs that were retrieved.

 ACCEPTED MANUSCRIPT
2.4 Data synthesis / analysis

Studies were grouped according to study design and purpose, such as case

studies, effectiveness studies, prognostic, prevalence and cross-sectional

studies; and summarised and tabulated accordingly. A narrative summary was

PT
used mostly, except regarding prevalence, for which a meta-analysis was

RI
conducted, pooling individual studies for totals regarding Centralization,

SC
Directional Preference, and no Directional Preference.

U
3. Results
AN
3.1.1 Study selection and characteristics of studies
M

2486 titles and abstracts were initially screened, 101 full texts were reviewed
D

for eligibility, and 43 articles were finally included (see Figure 1). The 43
TE

additional studies since the last review (May and Aina 2012) were: randomized
EP

controlled trials or controlled trials (10), or their secondary analyses (4), cohort

studies (15), or case studies (10), and four cross-sectional studies (see Table 1
C

for full details). Seven papers related to patients with neck pain; the rest to
AC

patients with low back pain. Most studies involved patients with non-specific

neck or back pain; but studies also included specific pain syndromes, including:

sciatica or cervical radiculopathy (3), discogenic pain (2), candidates for lumbar

disc surgery (2), potential red flags (1), and spinal stenosis (1). Although
 ACCEPTED MANUSCRIPT
centralization and DP were originally concepts related to the McKenzie

approach of MDT, other classification systems were also referred to, which all

involved some element of these clinical responses. Specifically these were the

Treatment-Based classification system (Heinz and Hegedus, 2008; Stanton et

PT
al., 2011), the diagnosis-based clinical decision guide (Murphy and Hurwitz,

RI
20111, 2), the Hall classification system (Gregg et al., 2014), a combined

SC
McKenzie and patho-anatomical assessment (Flavell et al., 2016), and a

discogenic sub-group from a wider classification system (Surkitt et al., 2016).

U
These studies will be discussed relative to their study designs and purpose in
AN
the results section (Table 1).
M

3.1.2 Case study designs

D

Surprisingly, given the length of time that the concepts of centralization and
TE

directional preference have been extant, ten studies were of a case study
EP

design or case series, which is generally considered the weakest of study

C

designs (Muir Gray, 1997). One case study (Desai et al., 2012), and two cohort
AC

studies (Van Helvoirt et al., 2014; 2016) described the effect of transforaminal

epidural steroid injections at reversing cervical or lumbar disc herniations with

radiculopathy from likely surgical candidates into potentially responding to

conservative interventions and demonstrating centralization or directional

preference. Other case studies demonstrated unusual presentations, such as

 ACCEPTED MANUSCRIPT
patients with spinal fractures, use of the treatment-based classification system,

spinal stenosis, disc displacement, a positive cranio-cervical flexion test, a

lateral component, or a patient with lower urinary tract symptoms. These

patients generally responded with centralization or directional preference,

PT
reversed these pathologies, and showed improvement over time (Heintz and

RI
Hegedus, 2008; Takasaki et al., 2010; Padmanabhan et al., 2011; Williams et

SC
al., 2011; Ojha et al., 2013; Elenburg et al., 2016; Robinson, 2016; Takasaki and

Herbowy, 2016; Wu and Rosedale, 2018).

U
AN
3.1.3 Effectiveness of exercises based on centralization or directional

preference
M

Some of the RCTs and trials of MDT utilising centralization and directional
D

preference (DP) demonstrated significant improvements in Global Perceived

TE

Effect and disability at two-three months, and one year (Petersen et al., 2011;
EP

Albert and Manniche, 2012; Halliday et al., 2016; Franz et al., 2017; ), and
C

disability at one month (Garcia et al. 2013) compared to a range of controls.

AC

But there were no significant differences in other trials (Bonnet et al., 2011;

Hosseinifar et al., 2013; Hagovska et al., 2014; Lopez-Diaz 2015; Moncelon

2015). In a retrospective analysis centralization or DP produced significantly

better function, but not pain in patients with neck pain compared to non-

centralizers (Edmond et al., 2014). Likewise in a small retrospective cohort of

 ACCEPTED MANUSCRIPT
patients with neck pain centralizers had better disability than non-centralizers

(Rose et al., 2016). In RCTs disability, but not pain, was significantly better

short-term (1m) compared to back school (Garcia et al., 2013); and also short-

term (8w) in Global Perceived Effect, but not other outcomes, compared to

PT
motor control exercises (Halliday et al., 2016); whereas motor control exercises

RI
had a better outcome short-term in another trial (Hosseinifar et al., 2013).

SC
3.1.4 Centralization and directional preference as treatment effect modifiers

U
Because of the nature of all study designs it was not possible to determine if
AN
either symptom response was a useful treatment effect modifier; no trial had

determined their presence at baseline, and then randomized patients to

M

management based on those concepts versus another management strategy.

D
TE

3.1.5 Prognosis of centralization and directional preference and other

prognostic factors
EP

Centralization or peripheralization was not associated with any particular type

C

of disc lesion, but both improved more than the no pain response group
AC

(Albert et al., 2012). Other secondary analyses looked at factors that improved

outcomes. Older age was associated with better outcomes in a MDT group

compared to a back school group (Garcia et al., 2016). Age, severity of leg pain,

pain distribution, nerve root involvement and centralization were not found to
 ACCEPTED MANUSCRIPT
be treatment effect modifiers favouring MDT over manipulation; however

nerve root involvement and peripheralization together did make the chance of

success greater especially for the MDT group (Petersen et al., 2015).

In the other cohort studies or secondary analysis of RCTs centralization or

PT
directional preference compared to their absence or to guideline-based advice,

RI
was associated with better pain and functional outcomes, but mostly only in

SC
the short to medium-term (Werneke et al., 2011; Al-Obaidi et al., 2013;

Edmond et al., 2014; Gregg et al. 2014; Rose et al., 2016; Surkitt et al., 2016;

U
AN
Werneke et al., 2018; Yarnbowicz et al. 2018); but did not add to predictive

factors in one study (Werneke et al., 2016). (See table 1 for details)
M

3.1.6 Prevalence of centralization and directional preference

D
TE

The occurrence of centralization and directional preference could be calculated

from 21 studies (Table 2, which also shows the summary data from the
EP

previous review). Out of 5135 spinal patients centralization occurred in 2028

C

(39.5%), and directional preference in 1321 (26%); neither centralization nor

AC

directional preference was reported in 1716 (33.5%), and only 70 patients with

LBP were not counted in one of these groups. The total included 720 patients

with neck pain in who the following was reported: centralization, 56%,

directional preference, 18%, and no directional preference, 26%. Centralization

was found in 44% of those with chronic low back pain in the 11 papers that
 ACCEPTED MANUSCRIPT
reported specifically on chronic, as opposed to mixed symptom duration, of

low back pain.

Five studies reported on the plane of movement of the directional preference;

which was predominantly extension (about 80%) in four of them; generally

PT
smaller proportions with lateral movements (mostly 10-14%); and less than

RI
10% for flexion (Table 2 for prevalence from previous and this review).

SC
3.1.6 Cross-sectional studies

U
Only one recent study considered the reliability of therapists to identify
AN
centralization, directional preference, and other aspects of the MDT
M

assessment process (Werneke et al. 2014). Reliability was generally weak, with
D

15 kappa values (k) all below 0.44 ; the level of training in MDT that the
TE

therapists had undertaken did not make any difference. Three judgements

were poor (k < 0.20), 10 fair(k = 0.21-0.40), and two moderate (k = 0.41-0.60)
EP

according to Altman (1991). Another study looked at the reliability and

C

prevalence levels of the treatment-based classification system, which included

AC

directional preference (Stanton et al., 2011). One study looked at the effect of

centralization and directional preference on tests for spinal control (Apeldoorn

et al., 2016). Two studies examined the prevalence levels of different

classification systems including MDT responses (Flavell et al., 2016; Mazzone et

al., 2016)
 ACCEPTED MANUSCRIPT
3.1.7 Quality of the studies

Two authors (SM, NR) independently rated the quality of the 10 RCTs against

the PEDro quality scale, and of 12 cohort studies against the quality scale

(Hartvigsen et al., 2015). There was 97% and 84% agreement between raters

PT
(85 / 88 agreements; 126 / 150 agreements) respectively. Kappa values

RI
between the two authors were respectively 0.92 and 0.78, indicating excellent

SC
to good levels of reliability in the two judgements (Altman, 1991).

U
Rated against the PEDro quality scale it was concluded that four RCTs were
AN
low, three RCTs were moderate and three RCTs were high quality (Table 3). It

was concluded that the quality of the 12 cohort studies were as follows: six
M

were low, one was moderate and five were high quality (Table 4). Regarding
D

the effect of quality, in the RCTs four of the six moderate and high quality trials
TE

had positive outcomes for the MDT groups. In the cohort studies two of the six
EP

moderate and high quality analyses had positive outcomes for the MDT
C

groups; but in the others a clear dichotomous comparison was not possible.
AC

4. Discussion

The present and previous review (May and Aina, 2012) bring together over 100

pieces of evidence about centralization and directional preference. This

probably make them the most referenced clinical responses exposed during
 ACCEPTED MANUSCRIPT
routine physical examination of specific and non-specific spinal patients. The

focuses in the previous review were on the definitions used for centralization,

the prevalence of centralization and directional preference, and their role as

prognostic indicators. There was limited evidence for them as treatment effect

PT
modifiers, variable evidence for the reliability of assessment of centralization,

RI
evidence for extension movement as the most common directional preference,

SC
and some evidence of a link between centralization and discogenic problems.

Some of the focuses in the present review were similar, but study designs were

U
different between the two reviews. The previous review contained 36 studies
AN
(N = 7113 patients) from which prevalence data could be extracted, whereas
M

this review contained 21 studies (N = 5135 patients). Between reviews the

D

prevalence of centralization was very similar, about 40%, whereas there was a
TE

marked decline in reporting of directional preference, from 70% down to 26%.

The previous review only reported directional preference in five studies,

EP

compared to 13 in the present review; so it might be suggested that the

C

present estimate is more robust. In the present review there was also a much
AC

clearer identification of the absence of centralization or directional preference,

with a lack of either symptom response occurring in about a third of all

patients. This has important clinical implications; the point of centralization

and directional preference is that they direct patient management. If these

 ACCEPTED MANUSCRIPT
symptom responses are missing in about a third of patients, then patient

management is seriously compromised. However centralization and directional

preference between them would appear to account for a sizeable proportion

(60-70%) of patients.

PT
In the previous review symptom duration was a definite determinant of

RI
centralization; with 77% prevalence in acute patients (N = 317), and about 40%

SC
in patients with chronic and mixed duration symptoms (4305). In the present

review very few patients with acute / sub-acute LBP were included (250); all

U
AN
the other studies reported mixed or chronic duration of symptoms. So it is not

possible to make any present judgement about the role of symptom duration
M

in the occurrence of these phenomena.

D

In the previous review centralization was generally associated with a good

TE

prognosis, and non-centralization with a poor prognosis, whereas directional

EP

preference had limited evidence. However, the latter had some evidence, and
C

centralization limited evidence as a treatment effect modifier. In the present

AC

review there was some evidence that centralization and directional preference

were positive prognostic indicators in eight out of nine studies, although only

short to medium-term. Due to lack of appropriate study designs it was not

possible to comment on either as treatment effect modifiers. There was

conflicting evidence from recent RCTs that MDT-management based on these

 ACCEPTED MANUSCRIPT
concepts lead to more successful outcomes than control groups; with five out

of ten trials either way.

In the previous review reliability of the MDT assessment process had been

evaluated by six studies, and found to be very variable, mostly from moderate

PT
to good (Altman, 1991). In this review only one high quality study (Werneke et

RI
al., 2014) had evaluated reliability of several components of the MDT

SC
assessment process, including centralization and directional preference, and

overwhelmingly found it fair at best. So, despite training, it appears that

U
AN
therapists are not reliable at classifying sub-groups that are purported to

determine management.
M

In this review as in previous literature the movement of directional preference

D

and centralization has been pretty consistent. Most patients appear to respond
TE

to extension forces, and far fewer proportions respond to flexion or lateral

EP

forces.
C

There are some differences from the previous review. There were 30 cohort
AC

studies or secondary reviews of such, 16 RCTs or secondary analysis of such,

seven criterion validity studies, six reliability, two surveys and one mini case

series - a total of 62 studies (May and Aina, 2012). In the present review there

were 43 studies, of which there were 15 cohort studies, 14 RCTs or related

studies, four cross-sectional studies, and ten case studies. It seems very
 ACCEPTED MANUSCRIPT
surprising that at this distance from the foundation and development of the

McKenzie Method that virtually a quarter of the recent published articles are

simply case studies, which have so limited a role in the development of

evidence-based physiotherapy. Furthermore it is also disturbing that reliability

PT
amongst practitioners is still so untrustworthy. If even trained therapists

RI
cannot agree, then this is a major problem, as MDT is a practitioner-led

SC
classification system that leads to management strategies.

In the previous review there were several studies that appeared to link

U
AN
centralization with discogenic pain, although heterogeneous definitions of

Centralization produced different results, very high levels of specificity were

M

clearly linked to non-Centralization (Laslett et al. 2005). In the present review

D

only one study attempted to explore this link, and found that type of disc
TE

lesions, such as, whether contained or extruded, were not associated with
EP

centralization or non-centralization responses, as might have been expected

(Albert et al., 2012). However this study explored abnormal morphology,

C
AC

whereas the earlier study used provocation discography (Laslett et al. 2005), a

much more direct way to establish a link between pathology and symptoms. In

addition there was one case series and two cohort studies that evaluated the

ability of transforaminal steroid injections to make disc herniations with

 ACCEPTED MANUSCRIPT
radiculopathy amenable to directional preference management (Desai et al.,

2012; van Helvoirt et al., 2014; 2016).

These studies that investigated the McKenzie intervention using centralization

or directional preference add further to the literature about the inconclusive

PT
benefit of MDT. There is some indication for their therapeutic value, but no

RI
further evidence that they might be treatment effect modifiers. However there

SC
is still reasonably good evidence that both are a positive prognostic sign. In

other words the recognition of these clinical responses at baseline is a good

U
AN
indicator of outcome, perhaps regardless of the applied management strategy.

Additionally in this review we evaluated the quality of studies that were

M

included; but unfortunately this provided limited further information,

D

suggesting possibly that MDT had some added value, but not with any clarity.
TE

5. Conclusion.
EP

This review has synthesised literature from 62 previous studies, but also
C

evaluated 43 additional studies. The importance of centralization and

AC

directional preference as prognostic factors is probably overwhelming;

whether they indicate a particular management pathway is not clear.

Centralization and directional preference are still very important clinical

indicators to monitor during the taking of patients' history and physical

 ACCEPTED MANUSCRIPT
examination. Although about a third of patients may demonstrate neither

clinical response, they are still common and important prognostic indicators.

References
Aina A, May S, Clare H. The centralization phenomenon of spinal symptoms - a

PT
systematic review. Manual Therapy 2004;9:134-143.
Albert H, Hauge E, Manniche C. Centralization in patients with sciatica: are pain

RI
responses to repeated movement and positioning associated with outcome or
types of disc lesions? Eur Spine J 2011;21:630-636.

SC
Albert H, Manniche C. The efficacy of systematic active conservative treatment
for patients with severe sciatica. A single-blind, randomized, clinical controlled
trial. Spine 2012;37:531-542.

U
AN
Al-Obaidi SM, Al-Sayegh NA, Nakhi HB, Skaria N. Effectiveness of McKenzie
interventions in chronic low back pain: A comparison based on centralization
phenomenon utilizing selected bio-behavioral and physical measures. Int J
M

Phys Med & Rehab 2013;1:4.

Altman DG. 1991. Practical Statistics for Medical Research. London: Chapman
D

& Hall.
TE

Apeldoorn AT, van Helvoirt H, Meihuizen H, Tempelman H, Vandeput D, Knol

DL, Kamper SJ, Ostelo RW. The influence of centralization and directional
EP

preference on spinal control in patients with nonspecific low back pain. J

Orthop Sports Phys Ther 2016;46:258-269.
C

Bonnet F, Monnet S, Otero J. Short-term effects of a treatment according to

the directional preference of low back pain patients. A randomized clinical
AC

trial. Kinesitherapie Revue 2011;112:51-59.

Chorti AG, Chortis AG, Strimpakos N, McCarthy CJ, Lamb SE. The prognostic
value of symptom responses in the conservative treatment of spinal pain. A
systematic review. Spine 2009;34:2686-2699.
Delitto A, Cibulka MT, Erhard RE, Bowling RW, Tenhula JA. Evidence for use of
an extension-mobilization category in acute low back syndrome: a prescriptive
validation pilot study. Physical Therapy 1993;;73:216-222.
 ACCEPTED MANUSCRIPT
Desai MJ, Padmanabhan G, Simbasivan A, Kamanga-Sollo GG, Dharmappa A.
Directional preference following epidural steroid injection in three patients
with acute cervical radiculopathy. Pain Practice 2012;13:559-565.

Edmond SL, Cutrone G, Werneke M, Ward J, Grigsby D, Weinberg J, Oswald W,

Oliver D, McGill T, Hart DL. Association between centralization and directional
preference and functional and pain outcomes in patients with neck pain. J

PT
Orthop Sports Phys Ther 2014;44:68-75.

Elenburg JL, Foley BS, Roberts K, Bayliss AJ. Utilization of mechanical diagnosis

RI
and therapy (MDT) for the treatment of lumbar in the presence of known
lumbar transverse process fractures: a case study. J Man Manip Ther

SC
2016;24:74-79.

Flavell CA, Gordon S, Marshman L. Classification characteristics of a chronic

U
low back pain population using a combined McKenzie and patho-anatomical
AN
assessment. Man Ther 2016;26:201-207.

Franz A, Lacasse A, Donelson R, Tousignant-Laflamme Y. Effectiveness of

M

directional preference to guide management of low back pain in Canadian

armed forces members: a pragmatic study. Mil Med 2017;182:e1957-e1966.
D

Garcia AN, Costa LdaCM, da Silva TM, Gondo FLB, Cyrillo FN, Costa RA, Costa
TE

LOP. Effectiveness of back school versus McKenzie exercises in patients with

chronic nonspecific low back pain: a randomized controlled trial. Phys Ther
2013;93:729-747.
EP

Garcia AN, Costa LdaCM, Hancock M, Costa LOP. Identifying patients with
chronic low back pain who respond best to Mechanical Diagnosis and Therapy:
C

secondary analysis of a randomized controlled trial. Physical Therapy

AC

2016;96:623-630.

Garner P, Hopewell S, Chandler J, MacLehose H, Akl EA, Beyene J et al., When

and how to update systematic reviews: consensus and checklist. Br Med J
2016;354:i3507.

Gregg CD, McIntosh G, Hall H, Hoffman CW. Prognostic factors associated with
low back pain outcomes. J Prim Health Care 2014;6:23-30.
 ACCEPTED MANUSCRIPT
Halliday MH, Pappas E, Hancock MJ, Clare HA, Pinto RZ, Robertson G, Ferreira
PH. A randomized controlled trial comparing the McKenzie Method to motor
control exercises in people with chronic low back pain and a directional
preference. J Orthop Sports Phys Ther 2016; 46:514-522.
Hagovska M, Takac P, Petrvicova J. Change in the muscle tension of erector
spinae after the application of the McKenzie Method in patients with chronic
low back pain. Phys Med Rehab Kuror 2014;24:133-140.

PT
Hartvigsen L, Kongsted A, Hestbaek L. Clinical examination findings as
prognostic factors in low back pain: a systematic review of the literature. Chiro

RI
& Man Ther 2015;23:13.
Hayden JA, van der Windt DA, Cartwright JL, Cote P, Bombardier C. Assessing

SC
bias in studies of prognostic factors. Ann Intern Med 2013;158:280-286.
Heintz MM, Hegedus EJ. Multimodal management of mechanical neck pain

U
using a treatment based classification system. J Man Manip Ther 2008;16:217-
224.
AN
Hosseinifar M, Akbari M, Behtash H, Amiri M, Sarrafzadeh J. The effects of
stabilization and McKenzie exercises on transverse abdominis and multifidus
M

muscle thickness, pain, and disability: a randomized controlled trial in

nonspecific chronic low back pain. J Phys Ther Sci 2013;25:1541-1545.
D

Laslett M, Oberg B, Aprill CN, McDonald B. Centralization as a predictor of

TE

provocation discography results in chronic low back pain, and the influence of
disability and distress on diagnostic power. Spine J 2005;5:370-380.
Long AL. The centralization phenomenon. Its usefulness as a predictor of
EP

outcome in conservative treatment of chronic low back pain (a pilot study).

Spine 1995;20:2513-2521.
C

Long A, Donelson R, Fung T. Does it matter which exercise? A randomized

AC

controlled trial of exercise for low back pain. Spine 2004;29:2593-2602.

Long A, May S, Fung T. The comparative prognostic value of directional
preference and centralization: a useful tool for front-line clinicians? J Manual
Manip Therapy 2009;16:248-254.
Lopez-Diaz JV, Arias-Buria JL, Lopez-Gordo E, Gordo SL, Oyarzun APA.
"Effectiveness of continuous vertebral resonant oscillation using the POLD
method in the treatment of lumbar disc hernia". A randomized controlled pilot
study. Man Ther 2015;20:481-486.
 ACCEPTED MANUSCRIPT
May S, Aina A. Centralization and directional preference: a systematic review.
Manual Therapy 2012;17:497-506.
Mazzone B, Wood R, Gombatto S. Spine kinematics during prone extension in
people with and without low back pain and among classification-specific low
back pain subgroups. J Orthop Sports Phys Ther 2016; 46:571-579.

McKenzie R, May S. The Lumbar Spine Mechanical Diagnosis and Therapy.

PT
Spinal Publications, New Zealand, 2003.
Moncelon S, Otero J. The McKenzie Method of mechanical diagnosis and

RI
Therapy in chronic low back pain with directional preference. Kinesither Rev
2015;15:31-37.

SC
Muir Gray JA. Evidence-based Healthcare. How to Make Health Policy and
Management Decisions. Churchill Livingstone, New York, 1997.

U
Murphy DR, Hurwitz EL1. Application of a diagnosis-based clinical decision
AN
guide in patients with low back pain. Chiro & Man Ther 2011;19:26.

Murphy DR, Hurwitz EL2. Application of a diagnosis-based clinical decision

M

guide in patients with neck pain. Chiro & Man Ther 2011;19:19.

Olivio SA, Macedo LG, Gadotti IC, Fuentes J, Stanton T, Magee DJ. Scales to
D

assess the quality of randomized controlled trials: a systematic review. Phys

TE

Ther 2008;88:156-175.
Ojha H, Egan W, Crane P. The addition of manipulation to an extension-
EP

oriented intervention for a patient with chronic LBP. J Man Manip Ther
2013;21:40-47.
C

Otero J, Bonnet F. Low back pain: prevalence of McKenzie's syndromes and

directional preference. Kinesither Rev 2014;14:36-44.
AC

Otero J, Bonnet F. Neck pain: prevalence of McKenzie's syndromes and

directional preference. Kinesither Rev 2016;16:2-10.

Padmanabhan G, Sambasivan A, Desai MJ. Three-step treadmill test and

McKenzie mechanical diagnosis and therapy to establish directional preference
in a patient with lumbar spinal stenosis: a case report. J Man Manip Ther
2011;19:35-41.
 ACCEPTED MANUSCRIPT
Petersen T, Larsen K, Nordsteen J, Olsen S, Fournier G, Jacobsen S. The
McKenzie method compared with manipulation when used adjunctive to
information and advice in low back pain patients presenting with centralization
and peripheralization. A randomized controlled trial. Spine 2011;36:1999-2010.

Petersen T, Christensen R, Juhl C. Predicting a clinically important outcome in

patients with low back pain following McKenzie therapy and spinal

PT
manipulation: a stratified analysis in a randomized controlled trial. BMC
Musculoskel Dis 2015;16:74.

RI
Robinson M. Clinical diagnosis and treatment of patient with low back pain
using the patient response model: a case report. Physio Theory Pract

SC
2016;32:315-323.

Rose T, Butler J, Salinas N, Stolfus R, Wheatley T, Schenk R. Measurement of

U
outcomes for patients with centralising versus non-centralising neck pain. J
AN
Man Manip Ther 2016;24:264-268.

Sanderson S, Tatt ID, Higgins JPT. Tools for assessing quality and susceptibility
M

to bias in observational studies in epidemiology: a systematic review and

annotated bibliography. Int J Epidemiol 2007;36:666-676.
D

Shojania KG, Sampson M, Ansari MT, Doucette S, Moher D. How quickly do

systematic reviews go out of date? A survival analysis. Ann Intern Med
TE

2007;147:224-233.
Skyte L, May S, Petersen P. Centralization: its prognostic value in patients with
EP

referred symptoms and sciatica. Spine 2005;30:E293-299.

Stanton TR, Fritz JM, Hancock MJ, Latimer J, Maher CG, Wand BM, Parent EC.
C

Evaluation of a treatment-based classification algorithm for low back pain: a

cross-sectional study. Phys Ther 2011;91:496-509.
AC

Surkitt LD, Ford JJ, Chan AYP, Richards MC, Slater SL, Pizzari T, Hahne AJ.
Effects of individualised directional preference management versus advice for
reducible discogenic pain: a pre-planned secondary analysis of a randomised
controlled trial. Man Ther 2016;25:69-80.

Takasaki H, May S, Fazey PJ, Hall T. Nucleus pulposus deformation following

application of mechanical diagnosis and therapy: a single case report with
magnetic resonance imaging. J Man Manip Ther 2010;18:153-158.
 ACCEPTED MANUSCRIPT
Takasaki H, Herbowy S. Immediate improvement in the cranio-cervical flexion
test associated with MDT-based interventions: a case report. J Man Manip
Ther 2016;24:285-292.

Van Helvoirt H, Apeldoorn AT, Knol DL, Arts MP, Kamper SJ, van Tulder MW,
Ostelo RW. Transforaminal epidural steroid injections influence Mechanical
Diagnosis and Therapy (MDT) pain response classification on candidates for

PT
lumbar herniated disc surgery. J Back Musculoskel Rehab 2016; 29:351-359.

Van Helvoirt H, Apeldoorn AT, Ostelo RW, Knol DL, Arts MP, Kamper SJ, van

RI
Tulder MW. Transforaminal epidural steroid injections followed by Mechanical
Diagnosis and Therapy to prevent surgery for lumbar disc herniation. Pain Med

SC
2014;15:1100-1108.

Werneke M, Hart D. Centralization phenomenon as a prognostic factor for

U
chronic pain or disability. Spine 2001;26:758-765.
AN
Werneke MW, Hart DL, George SZ, Deutscher D, Stratford PW. Change in
psychosocial distress associated with pain and functional status outcomes in
M

patients with lumbar impairments referred to physical therapy services. J

Orthop Sports Phys Ther 2011;41:969-980.
D

Werneke MW, Deutscher D, Hart DL, Stratford P, Ladin J, Weinberg J, Herbowy

S, Resnik L. McKenzie lumbar classification: inter-rater agreement by physical
TE

therapists with different levels of formal McKenzie postgraduate training.

Spine 2014;39:E182-E190.
EP

Werneke MW, Edmond S, Deutscher D, Ward J, Grigsby D, Young M, McGill T,

McClenahan B, Weinberg J, Davidow AL. Effecting of adding McKenzie
C

syndromes, centralization, directional preference, and psychosocial

AC

classification variables to a risk-adjusted model predicting functional status

outcomes for patients with lumbar impairments. J Orthop Sports Phys Ther
2016; 46:726-741.

Werneke MW, Edmond S, Grigsby D, McClenahan B, McGill T. Directional

preference and functional outcomes among subjects classified at high
psychosocial risk using STarT. Physio Res Int. 2018;e1711.
 ACCEPTED MANUSCRIPT
Williams B, Vaughn D, Holwerda T. A mechanical diagnosis and treatment
(MDT) approach for a patient with discogenic low back pain and a relevant
lateral component: a case report. J Man Manip Ther 2011;19:113-118.

Wu D, Rosedale R. The use of Mechanical Diagnosis and THerapy (MDT) in

patients with lower urinary tract symptoms (LUTS): case series. Physio Theory
Pract 2018;26:1-9.

PT
Yarnbowicz R, Tao M, Owens A, Wlodarski M, Dolutan J. Pain pattern
classification and directional preference are associated with clinical outcomes

RI
for patients with low back pain. J Man Manip Ther 2018;26:18-24.

SC
Young JL, Rhon DI, Cleland JA, Snodgrass SJ. The influence of exercise dosing on
outcomes in patients with knee disorders: a systematic review. J Orth Sports
Phys Ther 2018;48:146-161.

U
AN
M
D
TE
C EP
AC
 ACCEPTED MANUSCRIPT

Table 1. Description of studies into centralization and directional preference (N = 43 )

First author Study design Population Participants Intervention: Follow-up: Outcomes - Results - only

PT
and date / Purpose from which after MDT / DP & weeks (w), clinical or SD between
participants inclusion / control OR months MDT-related groups

RI
were exclusion classification (m), year reported
recruited criteria (y)

SC
Albert & Randomised 477 with 181 Symptom- 2m, 1y (93- Global DP: Global
Manniche controlled sciatica randomized guided 95%) RMDQ (<0.008); NR
2012 trial (RCT) referred to (acute- exercises (DP) + Leg pain some SD

U
back centre chronic) stabilization NR signs

AN
exercises .V. EQ-5D
sham exercises Sick leave

M
Albert 2012 Prospective See above 181: 165 MDT Ax 3m, 1y Cent. Cent. 25.5%
cohort: (91%) back Decrease 44%

D
2ndary RCT / & referred NB 16%

TE
types of disc pain Peripheralizat Non-Cent. 7%
lesions ion 8%
EP
related to ISQ Types disc
pain MRI lesions not
C

responses associated with

Cent. / non-
AC

Cent.
Al-Obaidi Prospective 297: 193 105 CLBP: MDT 5w, 10w Pain Cent. pain with
2013 cohort / Cent. eligible 62 Cent. / 43 Fear- activities
v partial Cent. partial Cent. avoidance (<0.001); NS
 ACCEPTED MANUSCRIPT

(PC) Disability overall pain

Physical
performance

PT
Apeldoorn Prospective LBP + leg pain 114 acute- MDT Ax Ax only DP with Cent. 51 (45%)
2016 cohort / test- chronic LBP DP no Cent. 23 (20%)

RI
retest No DP 40 (35%)
changes DP with Cent.

SC
spinal control better spinal
after Ax control (<0.02)

U
Bonnet 2011 RCT LBP 54 LBP MDT 1w Cent. 62% v 17%

AN
Guideline- (0.008)
based group Disability
Pain

M
Desai 2012 Case studies / NP with 3 acute- MDT Ax after 2w DP only after Full resolution
effect of ESI cervical chronic ESI ESI of symptoms

D
on DP radiculopathy

TE
Edmond Retrospective Convenience 328 NP Classified as: Discharge Function Cent. 40%
2014 cohort / sample with acute- Cent. + DP, Pain DP 70%
EP
FOTO data chronic Non-Cent + DP, Cent. or DP:
classified as Non-Cent + function
C

Cent. / non- non-DP (<0.01); pain

AC

Cent. (NS)
Elenburg Case study / Not given 1 LBP with MDT Ax & 1m Function Almost full
2016 MDT despite lumbar treatment Pain resolution
spine fractures
fractures
 ACCEPTED MANUSCRIPT

Flavell 2016 Prospective 316: 197 150 CLBP MDT Ax Ax only Cent. / Periph 32%
cross- (62%) (76%) Dysfunction 36%
sectional / Other 31%

PT
classification Postural 1%
systems

RI
Franz 2017 Pragmatic 47 consenting 44 military MDT (DP) (22) 3m Pain DP: pain, PGE,
controlled consecutive LBP Usual care (22) Disability disability 3m

SC
study LBP PGE (<0.05)
HC

U
Stability DP

AN
Garcia 2013 RCT 182 CLBP 148 CLBP MDT (DP) 1m, 3m, 6m Pain MDT: disability
(81%) Back school (BS) (99-100%) Disability 1m (0.004); all
QoL other (NS)

M
Garcia 2016 Prospective 140 / 148 Baseline 1m Pain Older age
cohort: (95%) DP characteristics Disability (0.01).

D
2ndary RCT / DP Cent. leg pain

TE
better high pain (NS)
responders
EP
DP
Gregg 2014 Retrospective Consecutive 1076 LBP Hall 6m Pain pain factors
C

cohort / classification* (<0.01); DP,

AC

factors and treatment. surgery (<0.09).

associated 12 prognostic Function age, shorter
with variables (age, pain (<0.001).
outcomes gender, pain, RTW job (<0.001),
disability, female, pain,
 ACCEPTED MANUSCRIPT

surgery, DP) DP (<0.07).

Hagovska Pragmatic Not given 31 LBP MDT 3m Pain NS between
2014 controlled discopathy Healthy Disability groups.

PT
study / effect 24 no-LBP controls Cent. Cent. 100% 1m
Cent. controls EMG

RI
Halliday 2016 RCT 133 70 (53%) MDT 8w Pain MDT: GPE
consented CLBP with MCE GPE (0.03)

SC
DP Function
Muscle

U
Heintz & Case study / Not given 1 NP TBC 6w Pain Cent. with

AN
Hegedus use of TBC Disability mobilisation
2008 ROM
Hosseinifar RCT 75:41 (55%) 37 (90%) MDT Discharge Pain

M
2013 MCE Disability MCE (<0.05)

D
Muscle
Lopez-Diaz RCT Not given 30 LBP Mobilization Discharge Pain Mobilization:

TE
2015 Modalities ROM Cent. (<0.001)
Cent.
EP
Mazzone Cross- Not given 18 LBP and Spine Ax only MSI CPR 100%
2016 sectional / 17 no LBP kinematics in subgroups:
C

kinematics LBP subgroups manipulation 35%

AC

during stabilisation 24%

extension Cent. 18%
DP 47%
Moncelon RCT Not given 14 CLBP with MDT Six sessions Function
2015 DP Usual care Pain
 ACCEPTED MANUSCRIPT

Murphy Prospective Consecutive 264 acute- According to Not Red flags 3%

20111 cohort / LBP in one chronic LBP classification recorded Cent. 41%
DBCDG year Pain 50%

PT
classification provocation
NR 24%

RI
Myofascial 10%
Murphy Prospective Consecutive 95 acute- According to Not Red flags 1%

SC
20112 cohort / NP in one chronic NP classification recorded Cent. 27%
DBCDG year Pain 69%

U
classification provocation

AN
NR 19%
Myofascial 22%
Ojha 2013 Case study / 2 Not given 1 CLBP TBC = DP + 7w Disability

M
categories manipulation ROM
TBC

D
Otero 2014 Prospective Consecutive 349 patients MDT Discharge Classification 92% Der.

TE
cohort / MDT with LBP Cent. 71% / 76% at
syndromes, discharge
EP
Cent. DP DP 73%
Stability MDT 90%
C

Otero 2016 Prospective Consecutive 297 patients MDT Discharge Classification 92% Der.
AC

cohort / MDT with NP Cent. 75% / 82% at

syndromes, discharge
Cent. DP DP 86%
Stability MDT 92%-
Padmana- Case study Not given 1 CLBP with DP + treadmill 3w Pain
 ACCEPTED MANUSCRIPT

bhan 2011 spinal Disability

stenosis ROM
Petersen RCT 1619: 350 350 CLBP MDT 2m, 1y Disability MDT: 2m, 1y

PT
2011 (27%) Cent. or Manipulation (93%) Pain Disability (0.02,
Peripheral- GPE 0.03).

RI
ization QoL Cent./Periph.
Satisfaction (NS)

SC
Further HC
Petersen 2ndary RCT / Not given 350 LBP as above Age MDT: NR +

U
2015 factors effect Duration Periph. (RR

AN
related to modifiers Pain variables 10.5)
positive NR
outcome in

M
RCT above
Robinson Case study Not given 1 sub-acute DP 4w Pain

D
2016 with DP LBP Disability

TE
ROM
MRI
EP
Rose 2016 Retrospective Not given 11 NP Cent. (6) Discharge Disability Cent: Disability:
cohort / Cent. Non-Cent (5) Cent. (0.005)
C

v non-Cent.
AC

Stanton 2011 Cross- 545 LBP > 90 250 acute or Testing out Ax only Manipulation 42%
sectional days subacute algorithm Stabilization 17.5%
study / LBP criteria for 4 DP 31%
Prevalence & sub-groups Traction: 9.5%
reliability TBC + 1 subgroup 25%
 ACCEPTED MANUSCRIPT

Kappa 0.52
Surkitt 2016 2ndary RCT 2038 CLBP + 78 met DP v Guideline- 5w, 10w, Pain DP: back pain
(Ford et al., leg pain DP + criteria based advice 26w, 52w Function 10w (0.003)

PT
2016) / Discogenic* Psychosocial
discogenic* General

RI
sub-group health
Takasaki Case study / Not given 1 LBP with MDT 1m Pain Cent. & disc

SC
2010 effect on disc MRI Disability displacement
MRI resolved

U
Takasaki Case study / NP 1 NP MDT Discharge CCFT CCFT negative

AN
2016 effect on after Cent.
CCFT
van Helvoirt Prospective 132 referred 69 non-Cent Transforaminal 2w, 1y Resolved 16%

M
2014 cohort / for HLDS HLDS epidural steroid Cent. 46%

D
effect of TESI candidates injection (TESI) Non-Cent / B 16%
Surgery 22%

TE
van Helvoirt Prospective 132 referred 77 non-Cent. TESI 1y Leg pain* Surgery v non-
2016 cohort / for HLDS HLDS Disability* surgery *
EP
2ndary above candidates GPE (0.001); Cent. v
different Back pain^ non-Cent *^
C

outcomes HADS^ (<0.05)

AC

Werneke Prospective Selected from 692 acute- MDT 1m Pain Non-Cent. v

2011 cohort / FOTO chronic LBP Functional Cent. worse
effect of Cent. database + leg pain status outcomes
on outcomes Psychological (<0.001)
distress
 ACCEPTED MANUSCRIPT

Werneke Reliability PT different 47 PTs 2 independent Ax only Agreement: Kappa 0.11 to

2014 study at levels levels of MDT 1662 MDT (MDT, DP, 0.44
MDT training training patients assessments Cent.)

PT
Level training
Werneke Retrospective 2066 LBP 723 for who MDT 1m Pain Cent. and DP

RI
2016 cohort / MDT, selected from complete Functional added little to
Cent. DP as FOTO data status predicting

SC
prognostic database Psychological outcomes
factors distress

U
Werneke Prospective LBP high 138 LBP DP v non-DP Discharge Disability DP (65%)

AN
2018 cohort / DP & STarT risk Other variables Psychological disability (0.03)
STarT from FOTO (pain, function, distress
MDT training)

M
Williams Case study Not given 1 discogenic MDT 2m Disability Resolution with
2011 with lateral LBP Cent.

D
component

TE
Wu 2018 Case studies Not given 3 CLBP with MDT <2m Prostate Complete
DP with LUTS LUTS Symptom resolution
EP
Index
Yarnbowicz Prospective 1006 LBP 940 initial DP Cent. Discharge Pain Cent. 20%
C

2018 cohort / Cent. consecutive 639 full data No DP no-Cent. Function Non-Cent. 39%
AC

DP patients Not classifiable Prognosis NC 23%

prevalence, & (NC) DP Cent. pain &
outcome function
(<0.001)
 ACCEPTED MANUSCRIPT

2ndary = secondary analysis of previous study; Ax = assessment; CLBP = chronic low back pain; CCFT = Cranio-cervical flexion test; CPR = clinical prediction
rules; DBCDG = diagnosis-based clinical decision guide; Der. = Derangement; EMG = electromyography of erector spinae muscle activity; ESI = epidural
steroid injection; FOTO = Focus on Therapeutic Outcomes; GPE = Global Perceived Effect; HADS = Hospital Anxiety Depression Scale; HC = healthcare;

PT
HLDS = herniated lumbar disc surgery; HE = healthcare; LUTS = lower urinary tract symptoms; MCE = motor control exercises; MDT = Mechanical
Diagnosis and Therapy or the McKenzie Method; MSI = movement system impairment; NR = nerve root; NS = no significant difference; PT = physical

RI
therapists; QoL = Quality of Life; RCT = randomized controlled trial; ROM = range of movement; RR = relative risk; STarT = subgroups for targeted
treatment back screening tool; TBC = treatment based classification system.

SC
*Discogenic = at least 4 of: Back +leg pain; sitting limited to 60 minutes; forward bending somewhat difficult; lifting somewhat difficult; sit to stand
somewhat difficult; coughing or sneezing somewhat difficult; symptoms worse the next day; working on manual job; onset associated with flexion /

U
rotation and/or compression loading.

AN
*Hall classification = four sub-groups based on site of symptoms and DP; and fifth sub-group with heightened pain behaviours (Gregg et al. 2014)

M
D
TE
C EP
AC
 ACCEPTED MANUSCRIPT

Table 2. Prevalence - Centralization and directional preference

Summary of previous studies - Centralization (N = 31) (May and Aina, 2012)

PT
Duration Symptoms N (%)

Acute LBP + NP 236 / 317 77%

RI
Sub-acute LBP 62 / 123 50%

SC
Chronic LBP 227 / 567 40%

U
Mixed LBP 1584 / 3738 42%

AN
Neck pain 62 / 168 37%

M
TOTAL 2109 / 4745 44%

Summary of previous studies - directional preference (N = 5), (May and Aina, 2012)

D
TE
TOTAL EP 1661 / 2368 70%

Studies from the present review (N = 21)

C

Duration Symptoms N Cent. DP No DP

AC

Albert (2012)N2 Mixed Sciatica 165 25% 59% 15%

Al-Obaidi (2013) Chronic LBP 105 59% 41%

Apeldoorn (2016) Mixed LBP +/- 114 45% 20% 35%

 ACCEPTED MANUSCRIPT

Edmond (2014) Mixed NP 328 40% 30% 30% (Ext. 80%, Flex. 10%, Lat. 10%)

Flavell (2016) Chronic LBP 150 32% 68%

PT
Garcia (2016) Chronic LBP +/- 148 95% 5% (Ext. 50%, Flex. 5%, Lat. 40%)

Hagovska (2014) Chronic Sciatica 31 100%

RI
Halliday (2016) Chronic LBP 133 73% 27% (Ext. 86%, Flex. 5.5%, Lat. 8.5%)

SC
Mazzone (2016) Chronic LBP 17 47% 53%

U
Murphy (2011)1 Chronic LBP +/- 264 41%

AN
Murphy (2011)2 Chronic NP +/- 95 27%

M
Otero (2014) Mixed LBP 349 76% 16% 8% (Ext. 80%, Flex. 4%, Lat. 13%)

Otero (2016) Mixed NP 297 82% 10% 8% (Ext. 84%, Flex. 3%, Lat. 14%)

D
TE
Petersen (2011) Chronic LBP +/- 350 53% 47%

Stanton (2011) (Sub)-Acute LBP +/- 250 31% 69%

EP
Surkitt (2016)N38 Chronic LBP +/- 78 51%
C

van Helvoirt (2014)N11 Chronic Sciatica post-TESI 69 46% 16% 22%

AC

Werneke (2011) Mixed LBP +/- 692 36/45% 64/55%*

Werneke (2016) Mixed LBP +/- 723 39% 29% 32%

Werneke (2018) Mixed LBP 138 65% 35%

 ACCEPTED MANUSCRIPT

Yarnbowicz (2018)N19 Mixed LBP 639 20% 64% 13%

* = depended on outcome: pain/function; Ext. = extension; Flex. = flexion; Lat. = Lateral

PT
N44
= missing numbers; the superscript number is the discrepancy between total and accounted for

RI
SC
Duration Symptoms N Cent. DP No DP UC

U
TOTAL Mixed LBP 2655 975 788 873 19

AN
Sciatica 265 104 108 40 13

M
Sub-acute LBP 250 77 173

Chronic LBP 1245 548 220 439 38

D
TE
Mixed Neck pain 720 401 128 191

TOTAL 5135 2028 1321 1716 70

EP
% 100% 39.5% 26% 33.5% 1%
C

Cent. = centralisation; DP = directional preference; no-DP = neither response; UC = uncounted

AC
 ACCEPTED MANUSCRIPT

Table 3. PEDro quality scale for randomised controlled trials (N = 10) (3 / 88 disagreements) 97% agreement

First 1 2 3 4 5 6 7 8 9 10 11 Total Overall

author and date out of 10 quality

PT
RI
Albert 2012 √ √ X √ X X
√ √ √ √ √ 7 Moderate
Bonnet 2011 √ √ X √ X X
X √ X √ X 4 Low

SC
Franz 2017 X X X √ X X
X √ X √ √ 5 Low
Garcia 2013 √ √ √ √ X X
√ √ √ √ √ 8 High

U
Hagovska 2014 X X X X X X
X √ X √ √ 3 Low

AN
Halliday 2016 √ √ √ √ X X
√ √ √ √ √ 8 High
Hosseinifar 2013 √ √ X √ X X
√ X X √ √ 5 Moderate

M
Lopez-Diez 2015 √ √ √ √ √ √ √ √ X √ √ 9 High
Moncelon 2015 √ X X X X X √ √ X √ √ 4 Low

D
Petersen 2011 √ √ √ √ X X √ X √ √ √ 7 Moderate

TE
PEDro scores: 1. Eligibility criteria were satisfied; 2. Subjects randomly allocated to groups; 3. Allocation was concealed; 4. Groups similar at baseline
EP
regarding most important prognostic indicators; 5. Blinding of subjects; 6. Blinding of all therapists; 7. Blinding of all assessors; 8. Measures of jey
outcomes were obtained from more than 85% of those initially allocated to groups; 9. All subjects for who outcome measures were available received the
C

treatment or the control as allocate, or where this was not the case, data were analysed by "intention to treat"; 10. Reports of between-group statistical
AC

comparison were reported for at least one key outcome; 11. Study provided both point measures and measures of variability for at least one outcome
measure. Score is out of ten item is not included.
 ACCEPTED MANUSCRIPT

Table 4. Prognostic study scores (N = 12) Disagreements 24 / 150 Agreement = 84%

PT
Author 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Total (15) Quality

RI
Albert 2012 √ √ √ √ √* √* √* √ √ √ √ √* √ √ √ 15 High
Al-Obaidi 2013 √ √ √ X X √ X √ √ √ √ √ √ X √ 11 High

SC
Edmond 2014 X X √ X X X √ √ √ √ √ X √ √ X 8 Moderate
Garcia 2016 √ √ √ √* √ √ √ X X √ √ √ √ √ √ 13 High
Gregg 2014 √ X √ X X X X √ X √ √ X X √ X 6 Low

U
Petersen 2015 √ √ √ X X √ √ √ X √ √ √ √ √ √ 12 High

AN
Rose 2016 X X X X √ √ √ X X √ √ X X X √ 6 Low
Surkitt 2016 √ √ X √ √ √ √ X X √ √ √ √ X √ 12 High
van Helvoirt 2014 √ √ √ √ √ √ X X X √ √ X X X X 8 Low

M
Werneke 2011 X X √ X X X X X X √ √ X √ √ √ 6 Low
Werneke 2016 X X √ X X X X √ √ √ √ X √ √ √ 8 Low

D
Werneke 2018 X X √ X X X X X X √ √ X √ √ √ 6 Low

TE
* with information from the accompanying study (Albert & Manniche 2012; Garcia et al. 2014; Petersen et al. 2011)

Quality items (from Hartvigsen et al. 2015): 1. Study population clearly defined; 2. Study population described; inclusion / exclusion criteria / chronicity; 3.
EP
Study population represent population of interest**; 4. Completeness of follow-up described for each point of follow-up to one year**; 5. Completeness
of follow-up adequate - 85%**; 6. Reasons for loss to follow-up adequately described; 7. No important differences (characteristics and outcomes)
C

between completers and non-completers; 8. Prognostic tests defined enough to be replicated; 9. Performance of prognostic tests are standardised; 10.
AC

Outcomes are defined; 11. Outcomes well established; 12. Method, setting, timing outcomes same for all participants; 13. Data presented sufficiently to
assess adequacy of analysis; 14. Statistical analysis sufficiently described and appropriate to account for other prognostic factors, such as multivariate
analysis**; 15. No selective reporting of results.
** these items were slightly amended from the original as described in the text.
 ACCEPTED MANUSCRIPT
Figure 1. Flow chart of selection process of studies

Screening titles and abstracts from

multiple database searching (N = 2487)

PT
Abstracts excluded (N = 2145)

RI
Duplicates removed (N = 255)

U SC
Additional records through
hand searching reference
AN
lists and mckenziemdt.org
website (N = 15)
M

Full-text articles assessed for

eligibility (N = 101) Full-text articles excluded (N = 58):
Extremity (15)
D

In previous systematic review (19)

Review or systematic review (8)
TE

Study protocol (3)

Inappropriate definitions used (5)
Letter / Comments only (8)
EP

Studies included (N = 43)

C
AC
 ACCEPTED MANUSCRIPT
• Centralization and Directional Preference are important clinical
findings
• This systematic review summarises 43 additional articles since 2012
• They are important prognostic indicators found in 60-70% of patients
• They should be routinely monitored in all spinal assessments
• Reliability was poor, and evidence they were treatment effect

PT
modifiers was absent.

RI
U SC
AN
M
D
TE
C EP
AC