REVIEW ARTICLE

Clinical course of non-specic low back pain: A systematic

review of prospective cohort studies set in primary care
C.J. Itz1, J.W. Geurts2, M. van Kleef2 P. Nelemans3
1 Department of Health Service Research, Maastricht University, and
Health Insurance Company VGZ, Eindhoven, The Netherlands
2 Department of Anaesthesiology and Pain Medicine, Maastricht University Medical Centre, 6202 AZ Maastricht, The Netherlands
3 Department of Epidemiology, Maastricht University, The Netherlands

Correspondence
Jose W. Geurts
E-mail: jose.geurts@mumc.nl
Funding sources
None declared.
Conicts of interest
None declared.
Accepted for publication
30 April 2012
doi:10.1002/j.1532-2149.2012.00170.x

Abstract
Background and objective: Non-specific low back pain is a relatively
common and recurrent condition for which at present there is no effective
cure. In current guidelines, the prognosis of acute non-specific back pain is
assumed to be favourable, but this assumption is mainly based on return to
function. This systematic review investigates the clinical course of pain in
patients with non-specific acute low back pain who seek treatment in
primary care.
Databases and data treatment: Included were prospective studies,
with follow-up of at least 12 months, that studied the prognosis of patients
with low back pain for less than 3 months of duration in primary care
settings. Proportions of patients still reporting pain during follow-up were
pooled using a random-effects model. Subgroup analyses were used to
identify sources of variation between the results of individual studies.
Results: A total of 11 studies were eligible for evaluation. In the first 3
months, recovery is observed in 33% of patients, but 1 year after onset,
65% still report pain. Subgroup analysis reveals that the pooled proportion
of patients still reporting pain after 1 year was 71% at 12 months for
studies that considered total absence of pain as a criterion for recovery
versus 57% for studies that used a less stringent definition. The pooled
proportion for Australian studies was 41% versus 69% for European or US
studies.
Conclusions: The findings of this review indicate that the assumption
that spontaneous recovery occurs in a large majority of patients is not
justified. There should be more focus on intensive follow-up of patients
who have not recovered within the first 3 months.

1. Introduction
Non-specific low back pain (LBP) is a relatively
common and recurrent condition with major medical
and economical implications for which today there is
no effective cure (van Tulder et al., 1995; Roelofs
et al., 2008; Becker et al., 2010; van Middelkoop
et al., 2011). Most treatment strategies and guidelines are based on the assumption that the prognosis
of acute LBP is favourable and that the pain resolves

spontaneously in the majority of patients (Spitzer

et al., 1987; Andersson, 1999; van Tulder et al.,
2006). However, the evidence for this statement is
mainly based on occupational studies in which
return to work or recovery from disability is
studied (Spitzer et al., 1987; Croft et al., 1998;
Andersson, 1999). These studies indicate that most
back pain patients return to function, this in spite of
their pain (Bowey-Morris et al., 2011). There seems
to be a lack of information on the course of acute

Eur J Pain 17 (2013) 515 2012 European Federation of International Association for the Study of Pain Chapters

Clinical course of non-specific low back pain

Database
This systematic review investigates the clinical
course of low back pain in primary care settings.
Nonspecific low back pain is a relatively common
and recurrent condition for which at present
there is no effective cure. In current guidelines
the prognosis of acute nonspecific back pain is
assumed to be favourable but this assumption is
mainly based on return to function. Included
were 11 prospective studies, with follow-up of at
least 12 months, which studied the prognosis of
patients with back pain for less than 3 months of
duration set in primary care.
What does this review add?
This review shows that after 3 months recovery
is observed in 33% and after one year 65% of
patients with low back pain still report pain.

non-specific LBP when pain rather than return to

work is considered as endpoint.
A previous systematic review that assessed the prognosis of acute LBP found high rates of LBP after 1 year
of follow-up (42% to 75%) (Hestbaek et al., 2003).
However, this aforementioned review did not evaluate
the clinical course by providing information on proportions of patients with early onset of LBP and also
included studies that were not performed in primary
care settings.
The present review is designed to investigate the
clinical course of pain in patients with non-specific LBP
of less than 3 months of duration, with a follow-up of at
least 12 months, and set in primary care.

2. Methods
2.1 Study selection
A literature search was performed for suitable articles
published between 1990 and 2010 in English, German
and Dutch, referenced on MEDLINE and PUBMED,
and EMBASE (Table 1). The search was started at 1990
because in 1987 Spitzer wrote his monograph: Scientific
approach to the assessment and management of activityrelated spinal disorders. A monograph for clinicians. Report
of the Quebec Task Force on Spinal Disorders (Spitzer et al.,
1987). This study had a major impact on the treatment
of LBP, and still has impact today. Therefore, we were
basically interested in evidence provided by studies
that were published in the years following this
publication.
6

C.J. Itz et al.

Table 1 Search strategy.

Search PUBMED
Search strategy:
Search (Low Back Pain[Majr] AND Follow Up Studies[Mesh]
AND Prognosis[Mesh]
NOT (trauma OR surgery OR children OR chronic)
[Title/Abstract]
Search MEDLINE
Search strategy:
1 low back pain OR sciatica (including related terms)
2 prognosis OR onset OR inception cohort OR follow up
(including related terms)
3 1 and 2
Search EMBASE
Search strategy:
Limit to human and years 19902010
Search terms used: back pain OR sciatica (title)
Follow up OR inception OR onset OR prognosis OR cohort
(abstract)
Acute OR subacute (abstract)

(350)
(87)

(10008)
(4823)
(24)

(29)

The following keywords were used: LBP and sciatica,

follow-up and prognosis, onset or inception cohort,
acute or sub-acute. Two authors (C.I. and J.G.) independently screened the titles, abstracts and keywords
of all references identified by the literature search to
determine if they addressed the research question.
Full-text publications were retrieved for potentially
relevant articles. The bibliographies of the retrieved
articles were screened for additional relevant papers.
Studies were considered eligible for review if they
allowed for evaluation of the clinical course of nonspecific LBP and met the following inclusion criteria:
(1) the study was prospective in design (prospective
cohort study or controlled trial); (2) the study population consisted of adult patients with non-specific
LBP; (3) patients were included within 3 months after
LBP onset with follow-up data of at least 1 year; (4)
one of the study outcomes was pain and the proportion of patients with or without pain could be
extracted from the study or could be established after
contacting the (corresponding) author; (5) the
patients were recruited in primary care settings; and
(6) data were available from patients who did not
undergo an intervention or from patients who underwent an intervention that was reported not to affect
the pain scores. Studies were excluded if: (1) the study
population included patients with trauma, surgery
and/or injury; and (2) the selection of patients was
restricted to special work conditions or pregnant
women. When multiple studies were identified with
overlap in study populations, only the original study
was included to avoid potential duplication of datasets.

Eur J Pain 17 (2013) 515 2012 European Federation of International Association for the Study of Pain Chapters

C.J. Itz et al.

Clinical course of non-specific low back pain

Table 2 Three methodological tests containing 12 individual criteria for prevalence studies to determine the quality of the studies.
1.
A.
B.
C.
2.
D.
E.
F.
G.
H.
I.
3.
J.
K.
L.

Is the nal sample representative of the target population?

At least one of the following must apply in the study: an entire target population, randomly selected sample, or sample stated to represent the
target population.
At least one of the following: reasons for non-response described, non-responders described, comparison of responders and non-responders,
or comparison of sample and target population.
Response rate and, if applicable, dropout rate reported.
Quality of the data?
Primary data of low back pain. (It was not taken from a survey not specically designed for that purpose.)
Appropriate use of statistics for the design of the study, and/or analysis described and appropriate, and/or analysis provides sufcient
presentation of data.
Same mode of data collection used for all subjects and, in longitudinal studies, at the rst and second data collection.
At least one of the following: questionnaire validated, tested for reproducibility, or tested (if low back pain information was collected by this
method).
At least one of the following: interview validated, tested for reproducibility, or adequately described and standardized (if low back pain
information was collected by this method).
At least one of the following: examination validated, tested for reproducibility, standardized or performed by the same person (if low back pain
information was collected by this method).
Denition of LBP
Precise anatomic delineation of the lumbar area or reference to an easily obtainable article that contains such specication.
Further useful specication of the denition of LBP, or question(s) put to study subjects quoted such as the frequency, duration or intensity, and
character of the pain, or reference to an easily obtainable article that contains such specication.
Recall periods clearly stated: e.g., 1 week, 1 month, or lifetime.

Modied Leboeuf criteria adding E: appropriate use of statistics (Leboeuf-Yde and Lauritsen, 1995). LBP, low back pain.

2.2 Data extraction

Two authors (C.I. and J.G.) independently extracted
data from selected studies on proportion of pain and
relevant population and study characteristics. The
main study parameter is the proportion of patients
with pain at 12 months. Other parameters of interest
were proportions of patients with pain at 1, 3 and 6
months. In cases where absolute numbers of patients
with pain at 12 months could not be derived from the
publication and/or the definition of recovery from
pain was unclear, the authors were contacted for additional information.
Study characteristics considered of interest were:
sample size; country where the study was performed;
year of publication; percentage of male participants;
mean age; definition and localization of LBP; mean
time since onset; and definition of recovery from pain.
Items concerning representativeness of the target
population such as response and dropout rates during
follow-up were also recorded.

2.3 Quality assessment

For assessment of the quality of the articles modified,
Leboeuf criteria were used (Leboeuf-Yde and Lauritsen, 1995) (Table 2). This method uses criteria related

to the representativeness of the study sample, the

quality of data and the definition of LBP. An additional
item concerning analysis of data (item E) was added.
The authors (C.I. and J.G.) independently scored these
items. In cases of disagreement, discrepancies were
discussed with a third author (P.N.) and consensus was
achieved. Each study was assigned a score, expressed
as a proportion of fulfilled criteria out of the total
number of relevant criteria. Information provided in
the published report of the study was scored as present
(+ criterion fulfilled), absent ( criterion not fulfilled)
or not applicable (NA) (Table 3). If the study design
appeared to allow for the omission of a certain criterion, it was noted as methodologically acceptable (+).
The main study parameter for this review was the
proportion of patients with pain. Therefore, if data on
pain were presented in a way that did not allow for
calculation of proportions, the score not applicable
was used for item E, otherwise it was scored as present
(+ criterion fulfilled). For each study, only one of the
items G, H and I was scored depending on the method
that was used to evaluate presence of pain (questionnaire, interview or examination).
We distinguished between studies with a quality
score of >70% versus studies with a score of 70% to
evaluate whether the quality of studies affects the
proportion of patients with pain after 1 year. The cutoff point of 70% was arbitrarily chosen.

Eur J Pain 17 (2013) 515 2012 European Federation of International Association for the Study of Pain Chapters

Clinical course of non-specific low back pain

C.J. Itz et al.

Table 3 Assessment of quality according to the modied Leboeuf criteria (Leboeuf-Yde and Lauritsen, 1995).
Representativeness

Quality of the data

Denition of low back pain

Study

Total %

(Bousema et al., 2007)

(Burton et al., 1999)
(Croft et al., 1998)
(Dettori et al., 1995)
(Epping-Jordan et al., 1998)
(Henschke et al., 2008)
(Klenerman et al., 1995)
(McGuirk et al., 2001)
(Schiottz-Christensen et al., 1999)
(Sieben et al., 2005)
(Werneke and Hart, 2001)

+
+
+
+
+
+
+
+
+

+
+
+
+
+
+
+
+

+
+
+
+
+
+
+
+
+
+

+
+
+
+
+
+

+
+
+
+
+
+
+
+
+
+
+

+
+
+
+
+
+
+
+

+
NA
+
+
+
+

NA
NA
+
NA
NA
NA
NA
NA
NA
NA
NA

NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA

+
+
+
+
-

+
+
+
+
-

+
+
+
+
+
+
+
+

70
60
50
50
80
90
40
80
60
80
80

+ Criterion fullled; - criterion not fullled; NA not applicable.

3. Results

2.4 Data analysis

The primary outcome of interest is the proportion of
patients who still suffer from LBP at 1 year after onset.
Secondary outcome measures were the proportion of
patients with LBP 1, 3 and 6 months after onset.
Proportions of patients with pain at 12 months, and
if available at 1, 3 and 6 months, were derived from
studies. For each time point, proportions were pooled
using random-effects models as proposed by DerSimonian and Laird using the inverse of the standard errors
of the proportion of individual studies as weights
(DerSimonian and Laird, 1986). The I2 index was used
to test for heterogeneity between study results. Significance of this index indicates that differences
between studies cannot be solely attributed to sampling variation and that differences in study population, design and analysis are responsible for variation
between study results (Higgins et al., 2003).
Subgroup analyses were used to evaluate whether
presence or absence of a specific study characteristic is
associated with higher or lower pooled proportions of
patients with pain at 12 months. For this purpose,
studies were categorized into subgroups according to
the presence or absence of a specific study characteristic. The differences in pooled proportions between
subgroups with corresponding 95% confidence intervals (CIs) were calculated to evaluate the magnitude
of effect of the study characteristic on study result and
to test the effect for statistical significance.
All analyses were performed with the statistical
package STATA (Copyright 2009, StataCorp LP, TX,
USA). P-values 0.05 were considered to indicate
statistical significance. Graphs were created with
either STATA or R (version 2.12.2: http://www.rproject.org/).
8

3.1 Study selection

The search strategy identified 99 papers eligible for
evaluation. After applying the inclusion and exclusion
criteria, 83 studies were excluded. Sixteen studies were
provisionally included for this systematic review
(Fig. 1). Two studies in which the population was a
subpopulation from an original study, which was
already included for this review, were excluded for
evaluation (Wahlgren et al., 1997; Costa Lda et al.,
2009). The authors of 10 studies were contacted by mail

Assessed for Eligibility n = 99

Not meeting inclusion criteria

n = 83

Provisional inclusion n = 16

Studies with subpopulation

of included studies
n=2

Authors contacted for additional pain data

Exclusion
n=3

Inclusion n = 11

Figure 1 Flow chart of the study selection.

Eur J Pain 17 (2013) 515 2012 European Federation of International Association for the Study of Pain Chapters

C.J. Itz et al.

and email to get additional information on proportions

of patients with pain at 1 year and, if available, other
follow-up time points and the exact definition used for
being pain free (Faas et al., 1993; Weber et al., 1993;
Dettori et al., 1995; Klenerman et al., 1995; Croft et al.,
1998; Epping-Jordan et al., 1998; Burton et al., 1999;
Werneke and Hart, 2001; Karjalainen et al., 2003;
Grotle et al., 2007). This approach resulted in a total of
11 studies that were finally included for evaluation in
this review (Dettori et al., 1995; Klenerman et al.,
1995; Croft et al., 1998; Epping-Jordan et al., 1998;
Burton et al., 1999; Schiottz-Christensen et al., 1999;
McGuirk et al., 2001; Werneke and Hart, 2001; Sieben
et al., 2005; Bousema et al., 2007; Henschke et al.,
2008).

3.2 Characteristics of included studies

In Table 4, the characteristics of the 11 included
studies are shown. The number of participants in each
study varied between 83 and 973. The percentage of
male participants varied between 45% and 100%. The
outcome parameter that was considered of primary
interest varied largely between the evaluated studies
and included pain but also physical activity and function; disability; fear avoidance; and sick leave. Two
studies were performed in the United States, two
studies in Australia and the remaining seven were
European studies.
The anatomical definition of LBP was mostly
defined according to localization of the pain; in six
studies no definition was stated. The definition of LBP
differs between studies, one study defined LBP as pain
in the thoracic and lumbar region, other studies formulated LBP as localized between the scapulae and
the gluteal folds, or below thoracic vertebra 6 (T6),
or between T12 and the buttock crease.
Different methods and pain scales were used for the
evaluation of pain intensity, namely: Visual Analogue
Scale (VAS); Numeric Rating Scale; Graded Chronic
Pain Scale; and Descriptor Differential Scale; a selfconstructed question about pain/a question about
pain specifically constructed for the study (Price et al.,
1983; Gracely and Kwilosz, 1988; Von Korff et al.,
1992; Childs et al., 2005).
Studies used different cut-off points for classifying
patients as free from pain and different periods in
which the patient had to be pain free (varying from 1
day to 6 months).
The mean time since onset of LBP varied between 0
and 12 weeks. In many studies, the timing of the
follow-up visits was at 1, 3, 6 and 12 months (Table 4).

Clinical course of non-specific low back pain

Methodological quality varied between 40% and 90%

(Table 3).

3.3 Clinical course of LBP

Fig. 2 shows the course of acute LBP during follow-up
of 1 year according to the pooled proportions of
patients with pain at 1, 3, 6 and 12 months. These
pooled proportions at 1, 3 and 6 months after onset
were 80% (95% CI: 61100%); 67% (95% CI:
5083%) and 57% (95% CI: 4668%), respectively.
The pooled proportion of patients with pain 1 year
after onset of LBP was 65% (95% CI: 5475%).
The Forest plot (Fig. 3) shows the proportions of
patients who still reported pain at 1 year after onset of
LBP with 95% CI for the individual 11 studies. The I2
index was 96.5 %, which indicates large heterogeneity
of study results.
There are five studies that reported proportions with
pain at both 3 and 12 months (Croft et al., 1998; Burton
et al., 1999; McGuirk et al., 2001; Sieben et al., 2005;
Henschke et al., 2008). All five studies showed that
after 3 months there was little additional recovery,
namely between 3 and 12 months the percentage of
patients still reporting pain decreased by only 17%.
Fig. 4 shows the results of subgroup analyses with
respect to the effect of pre-specified study characteristics on study results. The pooled proportion of patients
with pain at 1 year after onset was significantly lower
in two Australian studies than the pooled proportion
based on nine studies from Europe or the United
States. The pooled proportion of patients with pain at
1 year after onset was significantly lower for studies
that used a less stringent definition of recovery from
pain, i.e., studies that also considered patients who
reported mild pain as being free from pain and studies
that used a self-formulated question about pain/a
question about pain specifically formulated for the
study (Dworkin et al., 2005, 2009).

4. Discussion
The findings of this review indicate that the majority
of patients (65%) still experience pain 1 year after
onset of LBP. In the first 3 months, recovery is
observed in a substantial part of the patients, but
thereafter only few patients recover.
The conclusion of this review is in line with a previous systematic review that questioned the prognosis
of acute LBP and also found high rates of LBP after 1
year varying between 42% and 75% (Hestbaek et al.,
2003). This review differed from the present review by
also including studies that were performed in second-

Eur J Pain 17 (2013) 515 2012 European Federation of International Association for the Study of Pain Chapters

10

Patients from 35 GPs

Patients referred for

conservative treatment
to physiotherapists in
Virginia

Werneke and Hart

(2001)w United

Age in years

223

524

83

55

55

Pain intensity

Maximal pain
intensity

Pain/disability

Sick leave

Pain intensity

52

56

62

48

Fear-avoidance 50
behaviour

100

81

Pain
intensity

Functional
status

Disabling LBP

38 (10)

51% >45

38 (2946)

53 (4267)

43 (14)

32 (7)

28

43 (11)

47 (3852)

610

01 [ 3]

012 21(1463)

Acute pain localized

below scapulae and
above the gluteal folds
Not cervical

06

03

02 [4]

012 [14]

Pain localized below T12 02

and above the buttock
crease

01

Pain localized T6 or
below

54.7

25

30

25

52.7

20.8

Denition of pain
free at FU time point
(pain scale)

110/126
(87%)

Pain free in past week

(SF36; item 7)

573/955
(60%)

01 VAS (VAS 110) on

207/212
day of interview and free (98%)
of back-related disability
in the previous week
(question)
Free of recurrent pain at
612 months after onset
current LBP (6-point
scale)
Score of 0 (DDS 010)

3 weeks of LBP
complaints free and
current pain (VAS)
1 week of VAS
score of 0 (VAS 0100)

6 months (84%); 12
months (80%)

Score of 0 during past

week (NRS 010)

131/175
(75%)

37/72
(51%)

457/956
(48%)

149/188
(79%)

93/117
(79%)

133/191
(70%)

99/165
(60%)

291/484
(61%)

22/60
(37%)

107/177
(61%)

123/169
(72%)

235/491
(48%)

20/46
(44%)

97/123
(79%)

388/944
(41%)

63/78
(81%)

72/115
(63%)

128/170
(75%)

98/126
(78%)

62/89
(70%)

26 weeks 12 weeks 69 months 12 months

of pain (%) of pain (%) of pain (%) of pain (%)

Free of constant or
intermittent pain (pain
drawing)
3 months (87%); 6 months Score of 010 mm on
(72%); 12 months (55%)
pain intensity (VAS
0100 mm)
1 month (96%); 6 months Combination of sick
385/503
(92%); 12 months (94%)
leave and question about (77%)
feeling of well-being
concerning LBP (question
with answer yes)
3 months (81%); 6 months No back pain problem
(76%); 12 months (77%)
Score of 0 (GCPS 04)

12 months (41%)

6 weeks (99%); 3 months

(99%); 12 months (97%)

12 months (56%)

612 months (77%)

1 week (97%); 3 months

(86%); 12 months (78%)

2 weeks (77%); 12 weeks

(72%); 12 months (78%)

12 months (72%)

Onset LBP in
FU time points
weeks [mean
1LE- (response rate
days] median(IR) LBP % on pain scale)

Pain localized below

47 37(3245)
scapulae and above the
gluteal folds

012

Gender Mean (SD) Denition and

% male Median (IR) localization

Fear-avoidance 45
beliefs

Physical
activity

Main study
parameter

LBP, low back pain; , not specied; PCP, primary care practice; GP, general practitioner; FU, follow-up; 1LE-LBP, rst lifetime; IR, Interquartile Range; DDS, Descriptor Differential Scale (Gracely and Kwilosz, 1988); GCPS, Graded Chronic Pain Scale (Von Korff
et al., 1992); VAS, Visual Analogue Scale (Price et al., 1983); NRS, Numeric Rating Score (Childs et al., 2005); SF36, Quality of Life Short Form; SD, standard deviation.
a (Bousema et al., 2007); b (Burton et al., 1999); c (Croft et al., 1998); d (Dettori et al., 1995); e (Epping-Jordan et al., 1998); h (Henschke et al., 2008); k (Klenerman et al., 1995); m (McGuirk et al., 2001); s (Schiottz-Christensen et al., 1999); t (Sieben et al., 2005);
w (Werneke and Hart, 2001).

States of America
80%

Schiottz-Christensen
et al. (1999)s
Denmark 60%

Australia 80%

McGuirk et al. (2001)m

Klenerman et al. (1995)k 300

United Kingdom 40%

222
Sieben et al. (2005)t
The Netherlands 80%

USA Army active duty

personnel living in
Germany

Male patients from

the Naval Medical
Center in San Diego
Consecutive patients
from PCPs in Sydney
area
Patients from PCPs
Merseyside Region, i.e.,
metropolitan county
Control patients in a
comparative study,
patients from four GPs
Patients from 130 GPs in
North Jutland

149

Patients form PCPs

out of a previous
cross-sectional
population survey

140
Epping-Jordan et al.
(1998)e United States
of America 80%
Henschke et al. (2008)h 973
Australia 90%

Dettori et al. (1995)d

Germany 50%

162
North East United
Kingdom 60%
218
Croft et al. (1998)c
Manchester area
United Kingdom 50%

Burton et al. (1999)b

Patients from PCPs

and responders
advertisement
Patients from PCPs

124

Bousema et al. (2007)a

The Netherlands 70%

Baseline Description
N
of sample

Author (year), country

and quality score

Pain proportion

Current history/baseline FU

Characteristics of the study and social demographics

Denition of LBP

Main study parameter

Mean (SD) if not stated otherwise

Table 4 Details of the data extracted from the included studies.

Clinical course of non-specific low back pain

C.J. Itz et al.

Eur J Pain 17 (2013) 515 2012 European Federation of International Association for the Study of Pain Chapters

C.J. Itz et al.

Figure 2 Course of low back pain. Dots show pooled proportions. Error
bars show 95% condence intervals. The gures at the bottom of the gure
depict the number of studies that provided information for the specic
time points.

ary and tertiary care settings and no restriction to

recent onset of acute LBP. Despite the differences, both
reviews arrive at similar results. This finding may indicate that in the present review efforts to restrict the
study population to patients with early onset of back
pain have not been successful. The definition of recent
onset LBP is, with a duration of less than 3 months,
rather arbitrarily defined and relies heavily on the
memory of patients who may feel that their back pain
is of recent origin whereas it could have started more
than 3 months ago.
The findings in this review are in sharp contrast
with current recommendations and guidelines for the
treatment of patients with non-specific LBP, which
are based on the assumption that in a large majority
of patients spontaneous recovery occurs. The European guidelines for acute non-specific LBP cite that
acute LBP is usually self-limiting (a recovery rate of
90% within 6 weeks) and only 27% of people
develop chronic pain, although references to underpin this statement are not provided (van Tulder et al.,
2006). The assumption that spontaneous recovery is
common resulted in management recommendations
that put strong emphasis on reassurance of the
patient that rapid recovery is to be expected, limitation of referral to secondary care and continuation of
daily activities.
It may be worth considering what may be the basis
for this widespread belief that spontaneous recovery is
common. One of the reasons may be that in many
studies on back pain published during the last 20

Clinical course of non-specific low back pain

years, return to work or recovery from disability

was considered evidence for recovery from LBP
(Spitzer et al., 1987; Andersson, 1999). However, this
supposition may be criticized as it is quite conceivable
that patients may still suffer from pain. Another
reason may be that individual studies show variation
in results, although none of the reviewed studies
reported a recovery rate of 8090%.
Four larger studies that were included in this review
reported that the proportion of patients who are still
having pain 1 year after onset varied between 41% and
75%. Therefore, another aim of the present review was
to explore reasons for the large variation between
studies in pain results. An important source of heterogeneity that was identified was the definition of pain
recovery that was used. The subgroup of studies that
considered total absence of pain as a criterion for recovery and used a validated pain questionnaire, e.g., the
VAS, showed a higher pooled proportion of patients
with pain (71%) compared with the studies that used
less stringent standards and/or were content with considering low pain scores as indicative of complete
recovery (57%). The difference in the pooled proportions is 20% (Fig. 4).
Another interesting finding may be that studies performed in Australia (McGuirk et al., 2001; Henschke
et al., 2008) reported more favourable prognosis than
the studies from Europe and the United States. The
pooled proportion of patients with pain at 12 months
was lower in Australian studies than in American/
European studies, with a difference of 27% (Fig. 4).
One explanation for this finding could be that the
American/European studies generally used a combination of outcome measures regarding LBP. This is in
accordance with the IMMPACT recommendations by
Dworkin et al. who recommended use of a combination of relevant validated outcome measures to evaluate treatment effectiveness (Dworkin et al., 2005). In
the Australian study by McGuirk, only a VAS was used
and patients who reported mild pain, with one single
pain intensity score from 0 to 10 mm on a VAS from 0
to 100 mm, were considered as being free from pain.
The other Australian study by Henschke et al. used
only one modified question of the SF36 questionnaire
(Henschke et al., 2008) whereas the SF questionnaire
was not developed for this purpose.
The results of this and other systematic reviews indicate that the current approach towards management
of patients with non-specific LBP calls for reorientation. The paradigm that the prognosis of LBP is mostly
favourable can lead to conservatism in pain management and could be contra-productive for innovations
in pain treatment. It may have paralysed the need for

Eur J Pain 17 (2013) 515 2012 European Federation of International Association for the Study of Pain Chapters

11

Clinical course of non-specific low back pain

C.J. Itz et al.

Study
ID

ES (95% CI)

Weight

Bousema (2007)

0.70 (0.60, 0.79)

8.88

Burton (1999)

0.78 (0.71, 0.85)

9.18

Croft (1998)

0.75 (0.69, 0.82)

9.26

Dettori (1995)

0.63 (0.54, 0.71)

8.98

Epping (1998)

0.81 (0.72, 0.90)

8.99

Henschke (2008)

0.41 (0.38, 0.44)

9.53

Klenerman (1995)

0.79 (0.72, 0.86)

9.18

McGuirk (2001)

0.43 (0.29, 0.58)

8.11

Schiottz (1999)

0.46 (0.41, 0.50)

9.45

Sieben (2005)

0.73 (0.66, 0.79)

9.24

Werneke (2001)

0.63 (0.56, 0.70)

9.21

Overall (Isquared = 96.5%, p = 0.000)

0.65 (0.54, 0.75)

100.00

NOTE: Weights are from random effects analysis

Figure 3 Forest plot of a random-effects meta-analysis on the proportion of patients with low back pain 1 year after onset. The size of the square box is
proportional to the weight that each study contributes in the meta-analysis. The pooled estimate and 95% condence interval (CI) are marked by a
diamond.

knowledge about mechanism and causes of back pain

and hampered development of further treatment
options.
There should be more focus on intensive follow-up
and monitoring of patients who have not recovered
from pain within 3 months. Pharmacologic treatment and minimally invasive interventions must be
considered.
Further research is needed to re-evaluate the
concept of non-specific LBP. At present, LBP with
unknown cause is diagnosed as non-specific. But it
can not be excluded that within this heterogeneous
group identification of patients with specific causes is
possible. Classification into more specific subgroups
could result in more homogeneous groups and help
advance development of more pinpointed and specified pain treatment options.
12

This review has some limitations. First, results are

based on published data with a large variation in
study results. To account for this heterogeneity, a
random-effects model was used, but such a metaanalytic approach has limitations and therefore
results from pooling must be interpreted with
caution. However, if we had refrained from pooling,
the conclusion would still be that pain persists in a
substantial proportion of patients, as even studies
with conservative estimates indicate that at least
40% of patients are not free from pain after 1 year of
follow-up.
Second, for the evaluation of the course of LBP
over time, the pooled proportions at consecutive time
points were derived from different sets of studies.
There were only five studies that reported results at
both 3 and 12 months (Croft et al., 1998; Burton

Eur J Pain 17 (2013) 515 2012 European Federation of International Association for the Study of Pain Chapters

C.J. Itz et al.

Clinical course of non-specific low back pain

Difference in Study Characteristics

Difference in pooled proportions (95% CI)

0.06 (0.24, 0.12)

Quality of the study >70% versus <70%

0.02 (0.23, 0.26)

Sample size >100 vs <100

0.01 (0.20, 0.23)

Mean age in study >40 years vs <40 years

0.12 (0.06, 0.30)

European population vs American/Australian

0.05 (0.15, 0.26)

North American population vs European/Australian

Australian population vs European/American

0.27 (0.45, 0.10)

0.10 (0.27, 0.08)

Onset LBP <6 weeks vs >6 weeks

0.20 (0.06, 0.34)

Stringent definition of no pain vs less stringent definition

0.03 (0.16, 0.23)

Response rate >70% vs <70%

NOTE: Weights are from random effects analysis

Figure 4 Difference in pooled proportions with pain at 12 months between subgroups. Studies are categorized into subgroups according to presence
versus absence of a specic study characteristic. Presented are differences in pooled proportion with [95% condence intervals (CIs)] between subgroups
of studies. Positive difference indicates higher pooled proportion in studies in the rst subgroup compared with the pooled proportion of the second
subgroup. Negative difference indicates lower pooled proportion in the rst subgroup compared with the pooled proportion of the second subgroup.
LBP, low back pain.

et al., 1999; McGuirk et al., 2001; Sieben et al., 2005;

Henschke et al., 2008). The pooled proportions of
patients with pain from these studies were 67% (50
83%) and 62% (4481%) at 3 and 12 months,
respectively, and are consistent with the conclusion
that one-third of patients recover within the first 3
months and that the majority still report pain at 12
months.
Third, this review provides information on prevalence of pain at longer follow-up, but not on severity
of pain. Information on the distribution of pain
scores in patients with persisting pain was not provided in detail by the included studies, only one
study presented a mean VAS pain score of 26.5 mm
in patients who still suffer pain at 12 months of
follow-up (Bousema et al., 2007). It is recommended
to address this issue in more detail in future studies

on clinical course of patients with non-specific acute

LBP.

5. Conclusions
This systematic review shows that spontaneous recovery from non-specific LBP occurs in the first 3 months
after onset of LBP in about one-third of patients, but
the majority of patients (65%) still experience pain 1
year after onset of LBP. These findings indicate that the
assumption underlying current guidelines that spontaneous recovery occurs in a large majority of patients
is not justified. There should be more focus on intensive follow-up and monitoring of patients who have
not recovered within the first 3 months. Futureresearch should be directed at improvement of classification of non-specific LBP in more specific groups.

Eur J Pain 17 (2013) 515 2012 European Federation of International Association for the Study of Pain Chapters

13

Clinical course of non-specific low back pain

Author contributions
Both C.I. and J.G. independently screened the titles,
abstracts and keywords of all references identified by the
literature search, extracted data from selected studies on
population and study characteristics, and assessed the quality
of the articles. J.G. corresponded with authors from studies
considered for evaluation. Analyses were performed by J.G.
and P.N. P.N. and M.v.K. oversaw and contributed to the
overall execution of the project. All authors discussed the
results and commented on the manuscript. All authors
helped to write the manuscript.

Acknowledgement
The authors like to thank Sander van Kuijk from the Department of Epidemiology of the Maastricht University for
his help.

References
Andersson, G.B. (1999). Epidemiological features of chronic
low-back pain. Lancet 354, 581585.
Becker, A., Held, H., Redaelli, M., Strauch, K., Chenot, J.F.,
Leonhardt, C., Keller, S., Baum, E., Pfingsten, M.,
Hildebrandt, J., Basler, H.D., Kochen, M.M., DonnerBanzhoff, N. (2010). Low back pain in primary care: Costs of
care and prediction of future health care utilization. Spine 35,
17141720.
Bousema, E.J., Verbunt, J.A., Seelen, H.A., Vlaeyen, J.W., Knottnerus, J.A. (2007). Disuse and physical deconditioning in the
first year after the onset of back pain. Pain 130, 279286.
Bowey-Morris, J., Davis, S., Purcell-Jones, G., Watson, P.J.
(2011). Beliefs about back pain: Results of a population survey
of working age adults. Clin J Pain 27, 214224.
Burton, A.K., Waddell, G., Tillotson, K.M., Summerton, N.
(1999). Information and advice to patients with back pain can
have a positive effect. A randomized controlled trial of a novel
educational booklet in primary care. Spine 24, 24842491.
Childs, J.D., Piva, S.R., Fritz, J.M. (2005). Responsiveness of the
numeric pain rating scale in patients with low back pain. Spine
(Phila Pa 1976) 30, 13311334.
Costa Lda, C., Maher, C.G., McAuley, J.H., Hancock, M.J.,
Herbert, R.D., Refshauge, K.M., Henschke, N. (2009). Prognosis for patients with chronic low back pain: Inception cohort
study. BMJ 339, b3829.
Croft, P.R., Macfarlane, G.J., Papageorgiou, A.C., Thomas, E.,
Silman, A.J. (1998). Outcome of low back pain in general
practice: A prospective study. BMJ 316, 13561359.
DerSimonian, R., Laird, N. (1986). Meta-analysis in clinical
trials. Control Clin Trials 7, 177188.
Dettori, J.R., Bullock, S.H., Sutlive, T.G., Franklin, R.J., Patience,
T. (1995). The effects of spinal flexion and extension exercises
and their associated postures in patients with acute low back
pain. Spine 20, 23032312.
Dworkin, R.H., Turk, D.C., Farrar, J.T., Haythornthwaite, J.A.,
Jensen, M.P., Katz, N.P., Kerns, R.D., Stucki, G., Allen, R.R.,
Bellamy, N., Carr, D.B., Chandler, J., Cowan, P., Dionne, R.,
Galer, B.S., Hertz, S., Jadad, A.R., Kramer, L.D., Manning,
D.C., Martin, S., McCormick, C.G., McDermott, M.P.,

14

C.J. Itz et al.

McGrath, P., Quessy, S., Rappaport, B.A., Robbins, W.,

Robinson, J.P., Rothman, M., Royal, M.A., Simon, L., Stauffer,
J.W., Stein, W., Tollett, J., Wernicke, J., Witter, J. (2005). Core
outcome measures for chronic pain clinical trials: IMMPACT
recommendations. Pain 113, 919.
Dworkin, R.H., Turk, D.C., McDermott, M.P., Peirce-Sandner, S.,
Burke, L.B., Cowan, P., Farrar, J.T., Hertz, S., Raja, S.N.,
Rappaport, B.A., Rauschkolb, C., Sampaio, C. (2009). Interpreting the clinical importance of group differences in chronic
pain clinical trials: IMMPACT recommendations. Pain 146,
238244.
Epping-Jordan, J.E., Wahlgren, D.R., Williams, R.A., Pruitt, S.D.,
Slater, M.A., Patterson, T.L., Grant, I., Webster, J.S., Atkinson,
J.H. (1998). Transition to chronic pain in men with low back
pain: Predictive relationships among pain intensity, disability,
and depressive symptoms. Health Psychol 17, 421427.
Faas, A., Chavannes, A.W., van Eijk, J.T., Gubbels, J.W. (1993).
A randomized, placebo-controlled trial of exercise therapy in
patients with acute low back pain. Spine 18, 13881395.
Gracely, R.H., Kwilosz, D.M. (1988). The Descriptor Differential
Scale: Applying psychophysical principles to clinical pain
assessment. Pain 35, 279288.
Grotle, M., Brox, J.I., Glomsrod, B., Lonn, J.H., Vollestad, N.K.
(2007). Prognostic factors in first-time care seekers due to
acute low back pain. Eur J Pain 11, 290298.
Henschke, N., Maher, C.G., Refshauge, K.M., Herbert, R.D.,
Cumming, R.G., Bleasel, J., York, J., Das, A., McAuley, J.H.
(2008). Prognosis in patients with recent onset low back pain
in Australian primary care: Inception cohort study. BMJ 337,
a171.
Hestbaek, L., Leboeuf-Yde, C., Manniche, C. (2003). Low back
pain: What is the long-term course? A review of studies of
general patient populations. Eur Spine J 12, 149165.
Higgins, J.P., Thompson, S.G., Deeks, J.J., Altman, D.G. (2003).
Measuring inconsistency in meta-analyses. BMJ 327, 557
560.
Karjalainen, K., Malmivaara, A., Mutanen, P., Pohjolainen, T.,
Roine, R., Hurri, H. (2003). Outcome determinants of subacute low back pain. Spine 28, 26342640.
Klenerman, L., Slade, P.D., Stanley, I.M., Pennie, B., Reilly, J.P.,
Atchison, L.E., Troup, J.D., Rose, M.J. (1995). The prediction
of chronicity in patients with an acute attack of low back pain
in a general practice setting. Spine 20, 478484.
Leboeuf-Yde, C., Lauritsen, J.M. (1995). The prevalence of low
back pain in the literature. A structured review of 26 Nordic
studies from 1954 to 1993. Spine 20, 21122118.
McGuirk, B., King, W., Govind, J., Lowry, J., Bogduk, N. (2001).
Safety, efficacy, and cost effectiveness of evidence-based
guidelines for the management of acute low back pain in
primary care. Spine 26, 26152622.
Price, D.D., McGrath, P.A., Rafii, A., Buckingham, B. (1983). The
validation of visual analogue scales as ratio scale measures for
chronic and experimental pain. Pain 17, 4556.
Roelofs, P.D., Deyo, R.A., Koes, B.W., Scholten, R.J., van Tulder,
M.W. (2008). Nonsteroidal anti-inflammatory drugs for low
back pain: An updated Cochrane review. Spine 33, 17661774.
Schiottz-Christensen, B., Nielsen, G.L., Hansen, V.K., Schodt, T.,
Sorensen, H.T., Olesen, F. (1999). Long-term prognosis of
acute low back pain in patients seen in general practice: A
1-year prospective follow-up study. Fam Pract 16, 223232.
Sieben, J.M., Vlaeyen, J.W., Portegijs, P.J., Verbunt, J.A., van
Riet-Rutgers, S., Kester, A.D., Von Korff, M., Arntz, A., Knottnerus, J.A. (2005). A longitudinal study on the predictive

Eur J Pain 17 (2013) 515 2012 European Federation of International Association for the Study of Pain Chapters

C.J. Itz et al.

validity of the fear-avoidance model in low back pain. Pain

117, 162170.
Spitzer, W.O. (1987). Scientific approach to the assessment and
management of activity-related spinal disorders. A monograph for clinicians. Report of the Quebec Task Force on
Spinal Disorders. Spine 12, S159.
van Middelkoop, M., Rubinstein, S.M., Kuijpers, T., Verhagen,
A.P., Ostelo, R., Koes, B.W., van Tulder, M.W. (2011). A systematic review on the effectiveness of physical and rehabilitation interventions for chronic non-specific low back pain.
Eur Spine J 20, 1939.
van Tulder, M., Koes, B.W., Bouter, L.M. (1995). A cost-of-illness
study of back pain in The Netherlands. Pain 62, 233240.
van Tulder, M., Becker, A., Bekkering, T., Breen, A., Del Real,
M.T., Hutchinson, A., Koes, B., Laerum, E., Malmivaara, A.
(2006). Chapter 3. European guidelines for the management

Clinical course of non-specific low back pain

of acute nonspecific low back pain in primary care. Eur Spine

J 15(Suppl 2), S169S191.
Von Korff, M., Ormel, J., Keefe, F.J., Dworkin, S.F. (1992).
Grading the severity of chronic pain. Pain 50, 133149.
Wahlgren, D.R., Atkinson, J.H., Epping-Jordan, J.E., Williams,
R.A., Pruitt, S.D., Klapow, J.C., Patterson, T.L., Grant, I.,
Webster, J.S., Slater, M.A. (1997). One-year follow-up of first
onset low back pain. Pain 73, 213221.
Weber, H., Holme, I., Amlie, E. (1993). The natural course of
acute sciatica with nerve root symptoms in a double-blind
placebo-controlled trial evaluating the effect of piroxicam.
Spine 18, 14331438.
Werneke, M., Hart, D.L. (2001). Centralization phenomenon as
a prognostic factor for chronic low back pain and disability.
Spine 26, 758764; discussion 765.

Eur J Pain 17 (2013) 515 2012 European Federation of International Association for the Study of Pain Chapters

15