Spine (Phila Pa 1976). 2007 (Nov 1); 32 (23): 2630–2637
Jeffries LJ, Milanese SF, Grimmer-Somers KA.
Centre for Allied Health Evidence,
University of South Australia,
Adelaide, South Australia.
STUDY DESIGN: Systematic literature review.
OBJECTIVE: To explore the available research literature, and provide an up-to-date synthesis of the epidemiology of idiopathic adolescent spinal pain (IASP).
SUMMARY OF BACKGROUND DATA: IASP and its potential causes have been a concern to researchers for over 2 decades. Because it has been suggested that IASP is related to the incidence of adult spinal pain, it appears important to synthesize what is currently known about IASP.
METHOD: A systematic meta-synthesis approach was used to identify secondary review articles and primary epidemiological studies regarding any type of IASP (neck, upper back, or low back).
RESULTS: A total of 56 primary epidemiological (cross-sectional or longitudinal) studies were identified. Spinal or back pain was the most commonly reported measure, with the lifetime prevalence figures ranging from 4.7% to 74.4%. The lifetime prevalence of low back pain had a similar range, 7% to 72%. The prevalence of pain in other areas of the spine (i.e., thoracic spine and neck) was variably reported, as were incidence rates for all areas of the adolescent spine. IASP is thus a significant problem, and the prevalence figures approach those of adults. There is some evidence that IASP is a risk factor for spinal pain in adulthood. However, there was considerable variation in how back pain was defined, the areas of the spine that were reported on, the manner in which data were collected and reported, thus preventing any significant comparisons of prevalence or incidence rates across studies.
CONCLUSION: Although there is wide discrepancy in the manner in which adolescent spinal pain is reported, it is evident that lifetime prevalence rates increase steadily with age and approximate adult levels by around the age of 18 years. There is an opportunity for further longitudinal research, with standardized methodology, to be undertaken that builds on the findings from this large group of studies.
From the FULL TEXT Article:
Since the 1980s, idiopathic adolescent spinal pain (IASP)
and its potential causes have been a concern to researchers.
It has been suggested that IASP is related to the incidence
of adult spinal pain, possibly by underpinning the
establishment of psychosocial patterns, postures, experiences,
attitudes, concepts, beliefs, and behaviors related
to recurrent pain events. [1, 2] Although the incidence of
depression, anxiety, drug addiction, and eating disorders
significantly increases during adolescence, and the risk of
death is much higher in this period than in any other
stage of life,  adolescence is generally considered a
healthy time period from a musculoskeletal perspective. 
Spinal pain, however, appears to be a “normal life experience”  for many young people.
Improving our understanding of IASP, including its
impact and potential causes, requires an understanding
of earlier research findings, in particular the methodologic
processes underpinning standardization of data
recording. This article explores the currently available
research literature regarding the epidemiology of IASP,
using a systematic meta-synthesis approach and provides
a synthesis of what is currently known about IASP prevalence
and incidence, with suggestions for standardization
of research methodology to facilitate comparison
and synthesis of the literature in the future.
The aim of the review was to explore the available research
literature, and provide an up-to-date synthesis of the epidemiology
of IASP, including the methodology underpinning the
Studies included in this review were secondary review articles
that used a systematic approach to data identification, and
primary epidemiological studies (either cross-sectional or longitudinal).
Any type of spinal pain was eligible for inclusion in
the review [neck, upper back, and low back pain (LBP)].
The search terms employed for this review
were adolescent (teenager, young person), spinal pain (headache,
neck pain, upper back pain, LBP), epidemiology, prevalence,
incidence, cross-sectional and longitudinal.
Sources of Data.
The library databases explored for this study
included Medline, CINAHL, and AMED.
A purpose-built data extraction sheet (Appendix
1, available online through Article Plus) recorded information
on the country of study location, the age range of subjects,
the location of data collection (school, home, etc.), type of
epidemiological study design (cross-sectional, longitudinal),
year(s) of data collection and the statistics reported (prevalence,
This search identified 3 distinct types of primary studies:
cross-sectional (N 43), longitudinal reporting
prevalence data only (N 6), and longitudinal reporting
incidence data (which could include prevalence
data) (N 6).
Countries of Origin or Ethnicity
The studies were predominantly conducted in the western
world as seen in Appendix 1 (available online
through Article Plus).
Method of Data Collection
Data on spinal pain were collected in most instances by
self-administered questionnaires. There was considerable
variation, both between and within studies, in the
methods used to administer questionnaires, as seen in
Appendix 1 (available online through Article Plus).
Description of Spinal Pain
Spinal pain was variably reported in terms of “LBP,” “thoracic,”
“upper back” or “dorsal pain,” “neck pain” (including
various combinations of these areas e.g., neck and
thoracic pain), or generic “back pain.” Where generic
“back pain” was measured, information was variably presented
with regard to the specific sections of the spine (Appendix
1, available online through Article Plus).
Most of the epidemiological studies on IASP sampled students
from local schools, although Iwamoto et al (2004) 
used a group of athletes. Seven studies [2, 6-11] used a population
approach to sampling by accessing national databases
with sample numbers ranging from 33511to 189, 8947 (Appendix
1, available online through Article Plus).
Age and Gender of “Adolescents”
All cross-sectional studies collected data from both genders,
with the exception of 2 studies (boys only  and
girls only ). Inconsistent data reporting in the articles
reviewed precluded comparison of prevalence data
across all studies, by age and gender (Appendix 1, available
online through Article Plus).
Prevalence data for IASP was provided by both crosssectional
and longitudinal studies (Appendix 1, available
online through Article Plus).
Spinal Pain Prevalence.
Table 1 presents data on the range
of spinal pain prevalence data found in the literature.
Spinal or back pain was most often reported as lifetime
prevalence, with figures ranging from 4.7%  to 74.4% 
and 78.2%,  although the latter study included headache
and stomach ache, in addition to back ache.
LBP was typically reported in terms of
lifetime prevalence (i.e., “have you ever suffered from
LBP?”) with figures ranging from 7%8 to 72%.  Prevalence
rates for LBP in the last year ranged from 7%8 to
50.8%,  as can be seen in Table 2.
We collated the prevalence figures for studies investigating
adolescent LBP. Studies with subjects younger
than 10 years of age, or who were vague in their description
of LBP were excluded. Prevalence figures from 29
studies were extracted to form a picture of the variability
of the adolescent LBP prevalence figures in the literature.
Figure 1 demonstrates the spread of LBP prevalence data
related to recall periods of up to 1 week, 1 month, and
lifetime. The spread of data in the 1 week and 1 month
data sets were less marked, (2.2%–38%, N 12, and
2.4%–44.5%,N10, respectively) than that of lifetime
prevalence (7%–84%, N 19).
Neck, Thoracic and Shoulder Pain Prevalence.
prevalence data for the neck, thoracic spine and shoulder
was variably reported (i.e., neck and thoracic spine, neck
and shoulder) as seen in Table 3. The lifetime prevalence
of “neck pain” ranged from 3%  to 8%.  Thoracic pain
lifetime prevalence varied from 9.5%  to 72%.  Because
of the paucity of literature pertaining to prevalence
figures for adolescent neck, thoracic, and shoulder pain,
it was not considered feasible to pool results to examine
the spread of data.
Spinal Pain Incidence.
Table 4 presents pain incidence
data for all regions of the spine.
Statistical Analyses in Longitudinal Studies.
A number of
studies used odds ratios (OR) to test the relationship
between pain reports at baseline to pain reports at follow-
up. Harreby et al,  in a 25-year longitudinal study,
found that genetic factors and LBP in the adolescent period
were significantly associated with adult LBP (OR
2.81, P 0.00; OR 2.23, P 0.03). Hestbaek et al 
reported a significant increase in the probability of having
LBP in the previous year [OR 1.87, 95% confidence
interval (CI) 1.52–2.32], as well as persistent LBP (30
days in the previous year) (OR 1.43, 95% CI 1.18–1.73),
at 8 years follow-up in the case of LBP for more than 8
days in a year during adolescence. Siivola et al54 calculated
the relative risk of having weekly neck and shoulder
pain 7 years later (i.e., adults aged 22–25 years of age) as
significant (relative risk 2.5, 95% CI 1.1–2.8) for the
whole sample, and for females only, who had frequent
neck and shoulder symptoms during adolescence.
Summary of Findings
Tables 1 to 4, and Appendix 1 (available online through
Article Plus) provide a summary of the findings of this
literature review. The prevalence of IASP varied widely
across the literature, although there was a tendency for
prevalence and cumulative incidence of the various spinal
pains to increase as subjects age increased.
This systematic review of the literature from 1984 onward
confirms the high rates of childhood and adolescent
LBP found in an earlier review by Duggleby et al. 
Of concern is that a history of spinal pain has been reported
as a strong predictor of future spinal pain.  Because
the incidence of IASP is almost comparable to that
of the adult population by the end of adolescence,  it is
perhaps not surprising that various authors have postulated
on the likely link between adolescent and adult
spinal pain. [4, 11, 48]
Despite the plausible rationale that IASP could progress
to adult spinal pain, the literature is limited by small
numbers of longitudinal studies. Few studies [7, 11, 52, 54, 55]
have followed up subject samples for sufficiently long
periods of time. Of these studies, all except one  found
that the presence of spinal pain at baseline was a significant
risk for spinal pain in adulthood. Furthermore, onset
of pain early in life has been affirmed as predictive of
chronicity.  Thus, the evidence from this small number
of longitudinal studies lends support to the notion that
adolescent spinal pain is an important predictor of adult
Countries of Origin or Ethnicity
Only 3 studies, [32, 42, 50] focused on the epidemiology of
IASP outside the developed countries. Inferences about
the scope of this problem outside the developed countries, therefore, cannot be made. Only 1 study46 considered
the influence of ethnicity or race in pain reporting,
and found that the only significant difference was that 15-year-old black adolescents reported higher prevalence rates
of LBP than 15-year-old white adolescents (pp. 607).
Method of Data Collection
There was considerable variation in the method of data
collection, and this may have been a significant contributor
to the variation in prevalence and incidence figures
reported in the literature. Although most questionnaires
were self-administered, some studies used interviews [4, 20]
and others used a mixed method of interview and questionnaire [17, 30, 32, 40, 41]
or weekly information from
coaches about players with LBP with a follow-up physician
Questionnaires were most commonly administered at
school (24%) or at home (24%); however, some studies
(16%) administered the questionnaires to older students
at school, while younger students completed the questionnaires
at home with parental assistance. [2, 9, 16, 18, 21, 23, 26, 27]
It was often unclear to what extent parents assisted their
children in completion of the questionnaires at home.
The remaining studies did not clearly report location of
The process by which data are collected is an important
consideration when attempting to compare studies,
and even within a study a change in methodology from
self reported surveying to parental assisted reporting
may introduce the potential for differences in prevalence
rates to reflect an artifact of the varying methodology.9
It has been reported that although there is a correlation
between adolescent and parent-reporting of pain, it is not
recommended that they are used interchangeably. 
Description of Spinal Pain
Areas of the spine were often grouped in various combinations
(i.e., “spinal pain” or “thoracic and neck pain”),
making it difficult to extract the prevalence and incidence
of pain for each region of the spine. Differences in definitions
and delineations of pain made it difficult to contextualize
Lifetime prevalence and 1 year prevalence were the most
commonly reported measures of IASP. These measures are,
however, subject to “memory decay”  (p. 1749) where
episodes of pain are likely to be forgotten with the passage
of time, and “forward telescoping of events,”  where there
is a “tendency to recollect events such as back pain as having
happened more recently than they did” (p. 789). Support
for the influence of “memory decay” was found in a
Finnish study,  in which 26% of adult males who reported
sciatica in 1 questionnaire denied it 3 years later. Given this
significant rate of “memory decay” in adult subjects it is
interesting to hypothesize the potential influence on adolescents.
Burton et al  found a high level of forgetfulness of
previous LBP (12 month prevalence) in a group of 216
adolescents studied over the 5 years of their secondary
Almost 60% of the students who reported LBP
forgot at least 1 previous episode of spinal pain during
annual questioning. It could be hypothesized that the
greater spread of data in the longer recall periods (Figure 1)
is a reflection of influences such as memory decay. Given
this spread it appeared that the most reliable period for
recall in prevalence data collection is up to and including 1
month as it minimizes the potential biases of recollection.61
This provides further support for Staes et al’s  concerns
regarding the validity of longer term recall in the collection
of history of LBP data from adolescents. Staes et al  limited
their study recall period to 1month because it was assumed
that longer time recalls resulted in unreliable data. Similar
concerns regarding longer time recall periods have been
reported by Goodman and McGrath  (1 year prevalence)
and Hestbaek et al  (lifetime prevalence).
Most studies used a convenience sample made up of
school students; however, it is unclear whether children
who are absent or do not attend school are significantly
different (and in which way) from those who regularly
attend school. Samples of convenience further restrict the
extrapolation of results to the wider population.  National
population based surveys are often the best sampling
model as they facilitate large sample numbers,
eliminate potential biases from studying a discrete sample
group, and therefore enhance the applicability of results
across different population groups.
Age and Gender of “Adolescents”
The World Health Organization64 defines adolescence as
the period between 10 and 19 years of age. The United
Nations Population Fund  further divides the World
Health Organization definition into early adolescence
(10–14 years of age) and late adolescence (15–19 years
of age). Up until the age of 10 years of age the growth
and development of males and females is remarkably
similar.  Beyond the age of 10 years, as a by-product of
the process of puberty, the growth patterns of males and
females diverge significantly, both between genders, between
individuals and even within individuals.  Puberty
is considered to have stopped by at least 18 or 19 years of
age. Given the potential influence of puberty-directed
growth on the incidence or prevalence of adolescent spinal
pain48 it is important we ensure that the range of
pubertal ages are considered in epidemiological studies.
Most of the longitudinal studies included in this review
began collecting data from children during late primary
school or early high school age (range, 8–15 years)
and none of these studies considered or attempted to
establish the onset or duration of puberty.
Our aim, to provide an overview of the incidence or
prevalence of adolescent spinal pain was compromised
by a lack of comparability between the studies reviewed.
In particular we were frustrated by a lack of standardization
of definitions, data collection processes, and data
We believe that there is considerable scope for standardization
of methodologies to provide amore comprehensive
picture of the problem associated with IASP.
Currently there is no standard way in which spinal pain
is defined, collected, or described and thus comparison
between studies is constrained.
It seems however, despite significant variability in the
way that pain is reported, that there is strong evidence
that pain prevalence increases with age in the adolescent
period, usually slightly earlier in girls, possibly corresponding
with the time of puberty [1, 4, 7, 8, 14, 16, 27, 32, 36, 45, 46]
and there is some evidence that it is more common in
girls than boys. [7, 14, 16, 20, 26–28, 32, 36, 38, 49, 52, 54] The evidence
presented here also demonstrates that the lifetime
prevalence rates have reached adult levels by around the
age of 18 years. [4, 8, 16, 45, 57] Moreover, recent longitudinal
studies indicate that adolescent spinal pain, including the
“back,” “neck and shoulder,” and “low back,” in adolescence
is significantly associated with spinal pain in
adult life. There is an opportunity for further longitudinal
research to be undertaken that builds on the findings
from this large group of studies.
Where possible longitudinal studies should be carried
out to clearly demonstrate the changing prevalence
of IASP with age, and allow a better understanding
of the potential influences of pubertal
growth and environmental factors on the risk of
Standardization of period prevalence measures between
studies would help to ensure that studies are
comparable. We recommend that the duration of
the period prevalence should be minimized to ensure
validity of the data collected, by minimizing
the effects of “memory decay” and “forward telescoping
To reduce the potential for bias, and the possibility
that any differences in prevalence rates reflect differences
in methodology, the same process of data
collection (i.e., interviews or questionnaires administered
at school or at home) should be used for
all subjects within a sample.
Standardization of data reporting should be promoted
by the peer reviewers of submitted articles.
The reporting of prevalence or incidence data per
spinal area, age and gender will facilitate the process
of collating information about IASP in the
community. Data should be presented in numerical
format, not just as a graphs or bar charts, where
exact data are impossible to identify.
National population based surveys should be used
if at all possible to enhance the applicability of
results across different population groups.
It is recommended that either the full age range of
adolescents are used as subjects (i.e., 10–19 years
of age) or a measure of pubertal development is
included in the research methodology to reflect the
biologic age of the subject.
Idiopathic adolescent spinal pain (IASP) is a significant
public health issue, with prevalence figures
in adolescence approaching those in adults by
around 18 years of age.
IASP may be a risk factor for spinal pain as an
There is considerable variation in research methodology
with regard to IASP.