Content uploaded by Rahman Shiri
Author content
All content in this area was uploaded by Rahman Shiri on Mar 26, 2015
Content may be subject to copyright.
Systematic Reviews and Meta- and Pooled Analyses
Obesity as a Risk Factor for Sciatica: A Meta-Analysis
Rahman Shiri*, Tea Lallukka, Jaro Karppinen, and Eira Viikari-Juntura
*Correspondence to Dr. Rahman Shiri, Finnish Institute of Occupational Health, Topeliuksenkatu 41 a A, FI-00250 Helsinki, Finland
(e-mail: rahman.shiri@ttl.fi).
Initially submitted October 29, 2013; accepted for publication January 7, 2014.
The aim of this study was to assess the associations of overweight and obesity with lumbar radicular pain and
sciatica using a meta-analysis. We searched the PubMed, Embase, Scopus, and Web of Science databases from
1966 to July 2013. We performed a random-effects meta-analysis and assessed publication bias. We included 26
(8 cross-sectional, 7 case-control, and 11 cohort) studies. Both overweight (pooled odds ratio (OR) = 1.23, 95%
confidence interval (CI): 1.14, 1.33; n= 19,165) and obesity (OR = 1.40, 95% CI: 1.27, 1.55; n= 19,165) were
associated with lumbar radicular pain. The pooled odds ratio for physician-diagnosed sciaticawas 1.12 (95% CI: 1.04,
1.20; n= 109,724) for overweight and 1.31 (95% CI: 1.07, 1.62; n= 115,661) for obesity. Overweight (OR = 1.16,
95% CI: 1.09, 1.24; n= 358,328) and obesity (OR = 1.38, 95% CI: 1.23, 1.54; n= 358,328) were associated with
increased risk of hospitalization for sciatica, and overweight/obesity was associated with increased risk of surgery
for lumbar disc herniation (OR = 1.89, 95% CI: 1.25, 2.86; n= 73,982). Associations were similar for men and
women and were independent of the design and quality of included studies. There was no evidence of publication
bias. Our findings consistently showed that both overweight and obesity are risk factors for lumbar radicular pain
and sciatica in men and women, with a dose-response relationship.
back pain; hospitalization; intervertebral disc displacement; obesity; overweight; sciatica
Abbreviations: BMI, body mass index; CI, confidence interval; OR, odds ratio.
Low back pain is the number 1 debilitating condition glob-
ally, and in 2010 it contributed 10.7% to the total number of
years lived with disability (1). Among low back disorders,
sciatica and lumbar disc herniation are the most persistent
and disabling conditions (2). Lumbar radicular pain is de-
fined as low back pain radiating to the leg below the knee
level, while sciatica is defined as lumbar radicular pain along
with clinical findings suggestive of nerve root compression
(2,3). Sciatica is usually caused by compression or irritation
of one of the lumbosacral nerve roots, often due to derange-
ment of a lumbar intervertebral disc such as lumbar disc her-
niation (4), although other causes for sciatica have also been
reported (5).
The prevalence of lumbar radicular pain during the preced-
ing 12 months ranges between 13% and 36% (6–8) and that
of clinically defined sciatica between 2% and 5% (9–11). The
incidence of lumbar radicular pain increases with age, where-
as that of nonspecific low back pain tends to decrease with
age (12). Lumbar radicular pain and sciatica have poorer
prognoses than nonspecific low-back-pain syndromes, and
they can cause prolonged work disability (2,13).
The etiologies of lumbar radicular pain and sciatica are not
well known, but they seem to be multifactorial. Known risk
factors for lumbar radicular pain and sciatica include occupa-
tional workload, such as carrying heavy items, bending, or
kneeling (14,15), and body height (16,17). Lifestyle risk fac-
tors have also been suggested as possible risk factors for lum-
bar radicular pain and sciatica (18).
Obesity is a prevalent public health problem and is associ-
ated with various outcomes—in addition to cardiovascular
diseases, obesity has recently been associated with musculo-
skeletal disorders (18,19). Previously we conducted a sys-
tematic review on the associations of weight-related factors
with lumbar radicular pain and sciatica (18). Based on a qual-
itative assessment of 13 studies, we found associations of
weight-related factors with lumbar radicular pain or sciatica
929 Am J Epidemiol. 2014;179(8):929–937
American Journal of Epidemiology
© The Author 2014. Published by Oxford University Press on behalf of the Johns Hopkins Bloomberg School of
Public Health. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com.
Vol. 179, No. 8
DOI: 10.1093/aje/kwu007
Advance Access publication:
February 24, 2014
by guest on March 26, 2015http://aje.oxfordjournals.org/Downloaded from
in 1 out of 4 cross-sectional studies, 3 out of 4 case-control
studies, and 3 out of 5 cohort studies. Thus, so far it is un-
known whether only obesity is associated with lumbar radic-
ular pain and sciatica or both overweight and obesity are
associated. We also performed a meta-analysis on the rela-
tionships of overweight and obesity with nonspecificlow
back pain (19). Both overweight and obesity were associated
with an increased risk of nonspecific low back pain.
So far, the roles of overweight and obesity in lumbar radic-
ular pain or sciatica, as more specific or objectively assessed
outcomes, have not been addressed with a meta-analysis. Our
aim was to carryout a meta-analysis to estimate the magnitude
of the associations of overweight and obesity with lumbar ra-
dicular pain and/or sciatica. To include studies published after
our qualitative review, a period comprising the past 8 years, we
updated our search and reassessed the previous studies regard-
ing their eligibility for the meta-analysis.
METHODS
Search strategy
We conducted comprehensive literature searches in
PubMed, Embase, Scopus, and Web of Science using prede-
fined keywords (lumbar radicular pain or sciatic pain or sci-
atic syndrome or lumbosciatic syndrome or lumbosacral
radicular syndrome or sciatica or intervertebral disk displace-
ment or disc herniation or herniated lumbar disc or prolapsed
lumbar disc or disc protrusion or herniated nucleus pulposus or
spinal diseases or back pain or back disorders) and (BMI
or body mass index or overweight or underweight or obesity
or body weight or waist circumference or waist hip ratio). We
used both Medical Subject Headings and text words in
PubMed, and we used Emtree terms and text words in Em-
base. We included all languages, even though we did not
identify any eligible non-English paper. We excluded case re-
ports, reviews, guidelines, editorials, and letters. We checked
the reference lists of included articles for additional studies.
We looked at the full text of studies on the associations of
smoking and physical activity/inactivity with lumbar radicu-
lar pain or sciatica for additional studies on weight-related
factors (18). Moreover, we looked at the full text of studies
on the associations of overweight/obesity, smoking, and
physical activity/inactivity with low back pain to identify ad-
ditional studies on lumbar radicular pain or sciatica (19,20).
Selection of the studies
The first author (R.S.) assessed the titles, abstracts, and full
texts of the studies found and investigated whether the studies
examined the associations of weight-related factors with lum-
bar radicular pain or sciatica. We included cross-sectional
and cohort studies as well as both population-based and
hospital-based case-control studies in the systematic review.
To be eligible for a meta-analysis, the studies had to report
quantitative data on the association between overweight/obe-
sity and lumbar radicular pain or sciatica. We also contacted
several authors (16,21–25) for additional information or re-
sults. Some of them (16,22,25) provided additional informa-
tion or new results.
Quality assessment
Two reviewers (R.S. and T.L.) independently assessed the
quality of the studies using the Effective Public Health Prac-
tice Project tool for observational studies (26). Summary
quality scores may provide a useful overall assessment. How-
ever, the scales are not recommended for assessment of the
quality of studies in systematic reviews (27). Therefore, we
assessed 5 main domains: selection bias, performance bias,
detection bias, confounding, and attrition bias (see Web Table 1,
available at http://aje.oxfordjournals.org/). Studies conducted
among volunteers, studies that included patients with lumbar
radicular pain or sciatica without acontrol group, studies with
a response rate less than 50%, and studies not reporting quan-
titative results that could be used to estimate odds ratios were
excluded from the meta-analysis. Disagreements between the
2 reviewers were resolved by consensus.
Meta-analysis
We used World Health Organization recommended cutoff
points for body mass index (BMI; weight (kg)/height (m)
2
)
and defined overweight as BMI 25–29.9 and obesity as BMI
≥30 (28,29). We performed meta-analyses for overweight or
obesity and defined it as BMI ≥25. We also included studies
that reported an estimate for BMI ≥24 (30) or >24.3 (31) for
overweight/obesity or an estimate for BMI >27.5 (32), ≥28
(30,33), or ≥29 (31,34,35) for obesity. One study conducted
among adolescents defined overweight or obesity by using
internationally acceptable age-specific and sex-specific cut-
off points for BMI (36), and it was also included in the
meta-analysis.
For studies that analyzed BMI as a continuous variable
(1-unit increase in BMI), we estimated the effect size by
multiplying the log odds ratio by 5 for overweight and by
10 for obesity. For 1 study (37) that reported an estimate for
a 1-standard-deviation increase in BMI, we estimated the effect
size by dividing the log odds ratio by the standard deviation and
then multiplying by 5 for overweight and by 10 for obesity.
We pooled the estimates for the subgroups of BMI to ob-
tain an overall estimate for overweight or obesity. We also
pooled the estimates for subgroups of the study population
(e.g., men and women) to obtain an estimate for the total study
population. We calculated a new estimate for overweight or
obese subjects for studies that compared normal, overweight,
or obese people with underweight subjects (31,33–35). For
these studies, we calculated standard errors from the natural
logarithm of the confidence intervals, divided the relative
risk/odds ratio for overweight or obesity by the relative risk/
odds ratio for normal weight, and then estimated new confi-
dence intervals for the obtained relative risk/odds ratio.
For studies that reported mean BMI in participants with or
without sciatica, we calculated the standardized mean differ-
ence by dividing the difference between 2 mean values by the
pooled standard deviation. We then converted the standard-
ized mean difference to an odds ratio (38).
One cross-sectional study (39) did not report a confidence
interval for the estimate. We calculated the standard error
(SE) of the estimate from this study using the following for-
mula: SE = log(odds ratio)/Zvalue (40).
930 Shiri et al.
Am J Epidemiol. 2014;179(8):929–937
by guest on March 26, 2015http://aje.oxfordjournals.org/Downloaded from
We used a random-effects meta-analysis. The estimate
from a random-effects meta-analysis is more conservative
than that from a fixed-effect meta-analysis (40). Similar re-
sults were reported from cross-sectional, case-control, and
cohort studies (Table 1). Therefore, we combined all designs
in a single analysis. Four cohort studies reported the relative
risk of hospitalization or surgery due to lumbar disc hernia-
tion (16,33,35,41) for overweight or obese subjects. Since
the incidence rate of lumbar disc herniation was below 1% in
those studies, the odds ratios and relative risks were identical.
Thus, we did not convert the relative risks to odds ratios.
We assessed the presence of heterogeneity across the studies
by means of the I
2
statistic (42). The I
2
statistic shows the total
variation across studies that is not due to chance. An I
2
statistic
less than 25% indicates a small amount of inconsistency, and
more than 50% indicates a large amount of inconsistency (43).
We used meta-regression to determine whether study-level co-
variates accounted for the observed heterogeneity (44).
To assess publication bias, we used a funnel plot, which
compared the sizes of the overweight/obesity effects with
their standard errors. We used the Egger regression test to ex-
amine funnel plot asymmetry and the trim-and-fill method to
explore the number of missing studies due to publication bias
(45,46). Statistical significance for publication bias was
based on a Pvalue less than 0.10 (47). We used Stata, version
10 (StataCorp LP, College Station, Texas), for meta-analysis.
RESULTS
Our searches initially identified 5,303 abstracts (Web
Figure 1). The first author (R.S.) looked at the full text of
491 relevant study reports on the associations between weight-
related factors and low back pain or lumbar disc disorders.
There were 43 relevant studies on the associations of weight-
related factors with lumbar radicular pain or sciatica. We ex-
cluded 12 studies conducted among patient populations that
did not have a control group, 2 studies on volunteers, 2 stud-
ies with no quantitative data for estimation of the odds ratio,
and 1 study with a response rate of 40%. Finally, we included
8 cross-sectional studies, 7 case-control studies, and 11 co-
hort studies on the association between BMI and lumbar ra-
dicular pain or sciatica in the meta-analysis (Web Table 2).
Of the 26 studies included in this meta-analysis, 8 studies
were on lumbar radicular pain (6–8,25,30,34,37,48),
Table 1. Associations of Study Design, Sex, and Methodological Quality With the Size of the Relationship Between Overweight or Obesity and
Lumbar Radicular Pain or Sciatica (Sensitivity Analysis) in 26 Studies Included in a Meta-Analysis, 1966–2013
Study Characteristic
Overweight Overweight or Obesity Obesity
No. of
Studies OR 95% CI I
2
,% No. of
Studies OR 95% CI I
2
,% No. of
Studies OR 95% CI I
2
,%
Total 17 1.17 1.12, 1.22 0 25 1.32 1.19, 1.46 86.9 19 1.36 1.25, 1.48 32.2
Study design
Cross-sectional 5 1.15 1.06, 1.25 0 6 1.20 1.12, 1.28 1.6 7 1.37 1.21, 1.56 23.4
Case-control 3 1.15 1.02, 1.30 17.1 7 1.45 1.05, 2.00 95.4 3 1.34 1.17, 1.54 0
Cohort 9 1.19 1.12, 1.25 0 12 1.26 1.16, 1.38 47.7 9 1.40 1.18, 1.65 50.4
Sex of participants
Men 9 1.24 1.15, 1.33 0 13 1.26 1.18, 1.34 0 10 1.37 1.21, 1.54 0
Women 9 1.15 1.09, 1.22 0 13 1.23 1.14, 1.33 37.3 9 1.33 1.19, 1.49 36.6
Confounding
Weak 10 1.15 1.10, 1.21 0 12 1.19 1.13, 1.25 23.3 10 1.30 1.18, 1.43 30.9
Moderate/strong 7 1.23 1.12, 1.35 0 13 1.43 1.16, 1.77 90.6 9 1.47 1.28, 1.69 16.2
Selection bias
Weak 7 1.16 1.08, 1.25 22.0 11 1.47 1.18, 1.82 94.1 8 1.48 1.21, 1.82 69.0
Moderate/strong 10 1.18 1.11, 1.25 0 14 1.23 1.17, 1.28 0 11 1.36 1.25, 1.47 0
Performance bias
Weak 7 1.18 1.11, 1.25 0 10 1.39 1.12, 1.73 93.6 7 1.35 1.22, 1.50 0
Moderate/strong 10 1.16 1.09, 1.23 3.6 15 1.22 1.14, 1.31 37.6 12 1.40 1.23, 1.59 50.3
Detection bias
Weak 8 1.14 1.07, 1.21 0 16 1.34 1.13, 1.58 90.8 8 1.34 1.21, 1.48 0
Moderate/strong 9 1.19 1.12, 1.26 0 9 1.25 1.16, 1.36 45.9 11 1.41 1.24, 1.60 49.3
Attrition bias
Weak 11 1.16 1.10, 1.23 15.4 17 1.36 1.18, 1.56 91.2 13 1.38 1.23, 1.55 51.7
Moderate/strong 6 1.20 1.09, 1.32 0 8 1.25 1.16, 1.34 0 6 1.37 1.22, 1.55 0
Abbreviations: CI, confidence interval; OR, odds ratio.
Obesity and Sciatica 931
Am J Epidemiol. 2014;179(8):929–937
by guest on March 26, 2015http://aje.oxfordjournals.org/Downloaded from
7 were on clinically defined sciatica (9–11,35,39,49,50),
8 were on hospitalization due to sciatica (16,17,22,23,
31–33,51), and 4 were on surgery due to lumbar disc herni-
ation (22,24,41,52). One study (22) assessed the association
of overweight with hospitalization due to sciatica, as well as
surgery due to lumbar disc herniation. Of the 26 studies in-
cluded in the review, 17 studies reported results for over-
weight, 25 reported results for overweight/obesity, and 19
reported results for obesity.
Overweight/obesity and lumbar radicular pain or sciatica
The pooled odds ratio for lumbar radicular pain was 1.23
(95% confidence interval (CI): 1.14, 1.33; I
2
=0%; n=
19,165) for overweight and 1.40 (95% CI: 1.27, 1.55; I
2
=
0%; n= 19,165) for obesity (Figure 1). The pooled odds
ratio for physician-diagnosed sciatica was 1.12 (95% CI:
1.04, 1.20; I
2
=0%; n= 109,724) for overweight and 1.31
(95% CI: 1.07, 1.62; I
2
= 63.9%; n= 115,661) for obesity
(Figure 2).
The odds ratio for hospitalization due to sciatica was 1.16
(95% CI: 1.09, 1.24; I
2
=0%;n= 358,328) for overweight
and 1.38 (95% CI: 1.23, 1.54; I
2
=0%; n= 358,328) for
obesity (Figure 3). For surgery due to lumbar disc herniation,
the studies reported estimates only for overweight/obesity
(Figure 3). The pooled odds ratio was 1.89 (95% CI: 1.25,
2.86; I
2
= 79.7%; n= 73,982).
Heterogeneity and meta-regression
In this meta-analysis, the estimates of studies on the asso-
ciation between obesity and clinically defined sciatica were
moderately heterogeneous (I
2
= 63.9%; Figure 2), and those
of studies on the association between overweight/obesity and
surgery due to lumbar disc herniation were highly heteroge-
neous (I
2
= 79.7%; Figure 3).
Overweight
Riihimäki, 1989 (30)
Manninen, 1995 (37)
Lean, 1999 (6)
Miranda, 2001 (34)
Leino-Arjas, 2006 (48)
Kääriä, 2011 (7)
Karjalainen, 2013 (25)
Shiri, 2013 (8)
Subtotal (I2= 0.0%, P= 0. 773)
Overweight or Obesity
Riihimäki, 1989 (30)
Manninen, 1995 (37)
Lean, 1999 (6)
Miranda, 2001 (34)
Leino-Arjas, 2006 (48)
Kääriä, 2011 (7)
Shiri, 2013 (8)
Karjalainen, 2013 (25)
Subtotal (I2= 7.3%, P= 0. 374)
Obesity
Riihimäki, 1989 (30)
Manninen, 1995 (37)
Lean, 1999 (6)
Miranda, 2001 (34)
Leino-Arjas, 2006 (48)
Kääriä, 2011 (7)
Karjalainen, 2013 (25)
Shiri, 2013 (8)
Subtotal (I2= 0.0%, P= 0.759)
1.10 (0.60, 2.10)
0.96 (0.34, 2.77)
1.23 (1.10, 1.37)
1.08 (0.79, 1.38)
1.91 (1.11, 3.30)
1.27 (1.11, 1.45)
1.15 (0.80, 1.65)
1.30 (0.80, 1.90)
1.23 (1.14, 1.33)
1.31 (0.82, 2.07)
0.96 (0.38, 2.43)
1.26 (1.15, 1.38)
1.21 (0.98, 1.43)
2.12 (1.43, 3.16)
1.31 (1.17, 1.45)
1.40 (0.99, 1.98)
1.19 (0.86, 1.64)
1.29 (1.20, 1.38)
1.60 (0.80, 3.10)
0.93 (0.12, 7.68)
1.34 (1.15, 1.56)
1.42 (1.07, 1.76)
2.39 (1.34, 4.27)
1.38 (1.15, 1.66)
1.33 (0.65, 2.74)
1.60 (0.90, 2.80)
1.40 (1.27, 1.55)
ES (95% CI) Weight
1.10 (0.60, 2.10)
0.96 (0.34, 2.77)
1.23 (1.10, 1.37)
1.08 (0.79, 1.38)
1.91 (1.11, 3.30)
1.27 (1.11, 1.45)
1.15 (0.80, 1.65)
1.30 (0.80, 1.90)
1.23 (1.14, 1.33)
1.60 (0.80, 3.10)
0.93 (0.12, 7.68)
1.34 (1.15, 1.56)
1.42 (1.07, 1.76)
2.39 (1.34, 4.27)
1.38 (1.15, 1.66)
1.33 (0.65, 2.74)
1.60 (0.90, 2.80)
1.40 (1.27, 1.55)
, %
0.125 0.25 0.5 1.0 2.0 4.0 8.0
Odds Ratio
First Author, Year (Reference No.)
1.49
0.54
47.82
7.53
1.97
33.05
4.47
3.13
2.18
0.54
42.09
12.07
2.91
32.05
3.80
4.36
100.00
2.18
0.23
43.04
16.16
2.97
30.38
1.93
3.11
100.00
1.49
0.54
47.82
7.53
1.97
33.05
4.47
3.13
100.00
2.18
0.54
42.09
12.07
2.91
32.05
3.80
4.36
2.18
0.23
43.04
16.16
2.97
30.38
1.93
3.11
100.00
Figure 1. Results of a meta-analysis of the association of overweight or obesity with lumbar radicular pain, 1966–2013. The size of the gray
shaded area indicates the weight of each study. Horizontal lines show the 95% confidence intervals (CIs). ES, effect size.
932 Shiri et al.
Am J Epidemiol. 2014;179(8):929–937
by guest on March 26, 2015http://aje.oxfordjournals.org/Downloaded from
In the meta-regression of the 25 studies on the association
between overweight/obesity and lumbar radicular pain or
sciatica, the heterogeneity across studies was significantly
related to the type of outcome. The heterogeneity was not
explained by study design, selection bias, performance bias,
detection bias, attrition bias, or adjustment for potential con-
founders. Moreover, heterogeneity was not related to the use
of BMI as a continuous variable or the use of BMI as a cat-
egorical variable with 2, 3, or more categories.
After the exclusion of 1 study (11) from the meta-analysis on
the association between obesity and clinically defined sciatica,
the variation across studies disappeared and the I
2
statistic
dropped from 63.9% to 0%. For surgery due to lumbar disc her-
niation, the I
2
statistic also dropped from 79.7% to 0% when 2
case-control studies conducted in adults (24,52) (pooled OR =
2.57, 95% CI: 2.28, 2.88; I
2
= 0%) were analyzed separately
from 2 cohort studies conducted in adolescents (22,41) ( pooled
OR = 1.39, 95% CI: 1.03, 1.87; I
2
=0%).
Sensitivity analysis
In stratified analyses of all outcomes combined (Table 1),
the effect sizes were smaller in the studiesthat controlled their
estimates for potential confounders than in studies that re-
ported unadjusted estimates or controlled their estimates for
a few confounders only. In separate meta-analyses of differ-
ent outcomes, only the odds ratio for physician-diagnosed
sciatica among obese persons was attenuated in the studies
that controlled for potential confounders ( for overweight,
OR = 1.11, 95% CI: 1.03, 1.20 (I
2
= 0%); for obesity, OR =
1.13, 95% CI: 1.02, 1.24 (I
2
=0%)) (9,35). The estimates
were similar for lumbar radicular pain (overweight: OR =
1.22, 95% CI: 1.09, 1.36 (I
2
= 0%); obesity: OR = 1.39,
95% CI: 1.21, 1.61 (I
2
=0%))(7,25,34,37) and hospitaliza-
tion due to sciatica (overweight: OR = 1.16, 95% CI: 1.07,
1.27 (I
2
= 19%); obesity: OR =1.40, 95% CI: 1.22, 1.61 (I
2
=
14.3%)) (16,17,23,32).
The associations of both overweight and obesity with lum-
bar radicular pain or sciatica were similar in men and women.
The pooled odds ratio for obesity was 1.37 (95% CI: 1.21,
1.54) in men and 1.33 (95% CI: 1.19, 1.49) in women. The
effect sizes were similar according to study design and accord-
ing to the presence or absence of selection bias, performance
bias, detection bias, and attrition bias. Slight variation in the
effect sizes between different types of study designs was due
to the unequal distribution of the different types of outcomes.
Overweight
Heliövaara, 1991 (9)
Jhawar, 2006 (35)
Leino-Arjas, 2008 (10)
Subtotal (I2= 0.0%, P= 0. 598)
Overweight or Obesity
Heliövaara, 1991 (9)
Toda, 2000 (50)
Kostova, 2001 (49)
Jhawar, 2006 (35)
Leino-Arjas, 2008 (10)
Subtotal (I2= 0.0%, P= 0. 731)
Obesity
Heliövaara, 1991 (9)
Videman, 1995 (39)
Jhawar, 2006 (35)
Younes, 2006 (11)
Leino-Arjas, 2008 (10)
Subtotal (I2= 63.9%, P= 0.026)
1.05 (0.89, 1.18)
1.14 (1.04, 1.24)
1.20 (0.80, 1.79)
1.12 (1.04, 1.20)
1.06 (0.93, 1.20)
0.97 (0.60, 1.58)
1.32 (0.81, 2.18)
1.13 (1.05, 1.21)
1.24 (0.91, 1.68)
1.12 (1.06, 1.19)
1.10 (0.80, 1.40)
1.40 (1.00, 1.80)
1.13 (1.01, 1.25)
2.27 (1.48, 3.52)
1.29 (0.80, 2.08)
1.31 (1.07, 1.62)
Weight, %
1.06 (0.93, 1.20)
0.97 (0.60, 1.58)
1.32 (0.81, 2.18)
1.13 (1.05, 1.21)
1.24 (0.91, 1.68)
1.12 (1.06, 1.19)
1.10 (0.80, 1.40)
1.40 (1.00, 1.80)
1.13 (1.01, 1.25)
2.27 (1.48, 3.52)
1.29 (0.80, 2.08)
1.31 (1.07, 1.62)
1
0.125 0.25 0.5 1.0 2.0 4.0 8.0
Odds Ratio
ES (95% CI)
First Author, Year (Reference No.)
69.30
3.35
100.00
21.35
1.48
1.42
72.07
3.69
100.00
21.37
20.56
31.80
13.93
12.34
100.00
21.35
1.48
1.42
72.07
3.69
100.00
21.37
20.56
31.80
13.93
12.34
100.00
27.34
Figure 2. Results of a meta-analysis of the association of overweight or obesity with physician-diagnosed sciatica, 1966–2013. The size of the
gray shaded area indicates the weight of each study. Horizontal lines show the 95% confidence intervals (CIs). ES, effect size.
Obesity and Sciatica 933
Am J Epidemiol. 2014;179(8):929–937
by guest on March 26, 2015http://aje.oxfordjournals.org/Downloaded from
Publication bias
The pooled odds ratio from 26 studies on overweight/
obesity (7 studies) or obesity (19 studies) was 1.45 (95% CI:
1.26, 1.66). The funnel plot of data from the 26 studies in-
cluded in the meta-analysis was symmetrical (Figure 4). The
Pvalue for the Egger test was 0.107. No missing study due to
publication bias was imputed using the trim-and-fill method.
DISCUSSION
This meta-analysisshowed that both overweight andobesity
are consistently associated with an increased risk of lumbar ra-
dicular pain and sciatica, with a dose-response relationship
among both men and women.
We studied a wide range of outcomes, from self-reported
symptoms to objectively assessed outcomes such as clinically
verified sciatica with nerve root entrapment, hospitalization,
and surgery due to lumbar disc herniation. Although self-
reported radicular pain is a subjective outcome, we consid-
ered it clinically relevant, since it is associated with poorer
quality of life, more functional limitations, and increased
use of health-care services compared with nonspecificlow
back pain (53). We found consistent results for all outcomes.
The associations of overweight and obesity with lumbar ra-
dicular pain and sciatica were modest. The strengths of the
associations were similar to those for nonspecific low back
pain (19).
There were sex differences in the prevalence, incidence,
and recovery rates of our outcomes of interest; for example,
Overweight and Hospitalization
Heliövaara, 1987 (17)
Vessey, 1999 (33)
Kaila-Kangas, 2003 (32)
Leino-Arjas, 2004 (23)
Schumann , 2010 (31 )
Wahlström, 2012 (16)
Subtotal (I2= 0.0%, P= 0.440)
Over weigh t or Obesit y a nd H ospitaliz ation
Heliövaara, 1987 (17)
Vessey, 1999 (33)
Kaila-Kangas, 2003 (32)
Leino-Arjas, 2004 (23)
Zhang, 2009 (5 1)
Schumann , 2010 (31 )
Rivinoja, 2011 (22)
Wahlström, 2012 (16)
Subtotal (I2= 22.7%, P= 0.249)
Obesity and Hospitalization
Heliövaara, 1987 (17)
Vessey, 1999 (33)
Kaila-Kangas, 2003 (32)
Leino-Arjas, 2004 (23)
Schumann , 2010 (31 )
Wahlström, 2012 (16)
Subtotal (I2= 0.0% , P= 0.501)
Over weigh t or Obes ity and Surg er y
Böstman, 1993 (52)
Saftic, 2006 (24)
Mattila, 2 008 (41)
Rivinoja, 2011 (22)
Subtotal (I2= 79.7%, P= 0.002)
1.28 (0.99, 1.61)
1.02 (0.68, 1.35)
1.92 (0.84, 4.36)
1.08 (0.96, 1.22)
1.29 (0.98, 1.71)
1.19 (1.08, 1.31)
1.16 (1.09, 1.24)
1.33 (1.04, 1.62)
1.02 (0.73, 1.32)
2.25 (1.23, 4.10)
1.16 (1.05, 1.27)
1.12 (0.96, 1.29)
1.35 (1.08, 1.69)
1.13 (0.78, 1.64)
1.23 (1.13, 1.35)
1.20 (1.12, 1.29)
1.63 (0.99, 2.59)
1.04 (0.45, 1.63)
2.69 (1.12, 6.45)
1.30 (1.12, 1.52)
1.47 (1.00, 2.15)
1.45 (1.18, 1.78)
1.38 (1.23, 1.54)
2.56 (2.28, 2.88)
2.77 (1.05, 4.49)
1.52 (1.04, 2.21)
1.19 (0.73, 1.94)
1.89 (1.25, 2.86)
1.28 (0.99, 1.61)
1.02 (0.68, 1.35)
1.92 (0.84, 4.36)
1.08 (0.96, 1.22)
1.29 (0.98, 1.71)
1.19 (1.08, 1.31)
1.16 (1.09, 1.24)
1.33 (1.04, 1.62)
1.02 (0.73, 1.32)
2.25 (1.23, 4.10)
1.16 (1.05, 1.27)
1.12 (0.96, 1.29)
1.35 (1.08, 1.69)
1.13 (0.78, 1.64)
1.23 (1.13, 1.35)
1.20 (1.12, 1.29)
1.63 (0.99, 2.59)
1.04 (0.45, 1.63)
2.69 (1.12, 6.45)
1.30 (1.12, 1.52)
1.47 (1.00, 2.15)
1.45 (1.18, 1.78)
1.38 (1.23, 1.54)
2.56 (2.28, 2.88)
2.77 (1.05, 4.49)
1.52 (1.04, 2.21)
1.19 (0.73, 1.94)
1.89 (1.25, 2.86)
Weight, %
10.125 0.25 0.5 1.0 2.0 4.0 8.0
Odds Ratio
ES (95% CI)ES (95% CI)
33.37
7.81
3.93
0.68
32.17
5.84
49.57
100.00
8.30
4.95
1.28
28.05
15.89
8.00
3.25
30.28
100.00
5.31
2.96
1.60
52.65
8.42
29.05
100.00
16.72
26.73
23.18
100.00
First Author, Year (R eference No.)
Figure 3. Results of a meta-analysis of the associations of overweight, overweight/obesity, and obesity with hospitalization due to sciatica and of
the association of overweight/obesity with surgery due to sciatica, 1966–2013. The size of the grayshaded area indicates theweight of each study.
Horizontal lines show the 95% confidence intervals (CIs). ES, effect size.
934 Shiri et al.
Am J Epidemiol. 2014;179(8):929–937
by guest on March 26, 2015http://aje.oxfordjournals.org/Downloaded from
self-reported lumbar radicular pain is more prevalent among
women (25), and they seem to experience slower recovery
from severe sciatica (54), while men have a higher incidence
of hospitalization and surgery due to sciatica (22,55). How-
ever, our meta-analysis showed no sex difference in the asso-
ciation between overweight/obesity and lumbar radicular
pain or sciatica. A previous meta-analysis on the associations
between obesity and nonspecific low back pain showed stron-
ger associations for women than for men (19). Differences
between age groups in the associations of overweight and
obesity with lumbar radicular pain and sciatica may also
exist. However, the studies included in this meta-analysis
did not report any age-specific results.
The mechanisms by which obesity increases the risk of
lumbar radicular pain and sciatica are not known. Obesity
contributes to the development of chronic, low-grade inflam-
mation through release of inflammatory mediators from excess
adipose tissue (56). Obesity-related chronic inflammation may
lead to the developmentof sciatica or the persistence of sciatica
symptoms. Leptin is one of the adipocyte-derived adipokines,
and high serum leptin levels are associated with obesity and
with the development of knee osteoarthritis independently of
BMI (57). Leptin is suspected to be involved in reorganizing
the cytoskeleton of nucleus pulposus cells (58), but the role of
fat tissue-derived leptin in the association between obesity and
sciatica is not known.
Obesity may slow down the healing of a disc injury. In a
large trial of patients with sciatica (59), obese patients had less
improvement in their back-related disability than nonobese
patients. The slower recovery was observed irrespective of
the type of treatment (i.e., conservative or surgical). More-
over, obesity increases the risk of recurrent disc herniation
after lumbar microdiscectomy (60).
Obesity may also interfere with the nutrition of the inter-
vertebral discs, leading to an impaired healing process. In a
3-year follow-up study of sciatica patients (61), BMI was
the strongest predictor of incident lumbar artery occlusion,
which also suggests that impairment of nutrition can be one
of the pathways of obesity’s relationship with sciatica.
The results of any meta-analysis depend on the data from
the original studies, and publication bias can distort the findings.
There was no evidence of publication bias, however. Small
studies on lumbar radicular pain or sciatica with nonsignifi-
cant results for overweight/obesity have been published due
to the fact that the main aim in many of those studies was not
to examine the relationship of overweight/obesity with lum-
bar radicular pain or sciatica. Overweight/obesity in these
studies was used as a covariate for adjustment purposes. Fur-
thermore, the associations of overweight and obesity with
lumbar radicular pain or sciatica may have been under-
estimated in studies that did not use the World Health
Organization-recommended BMI cutoff points to define
overweight and obesity.
Our sensitivity analyses showed that the findings of this
meta-analysis are robust. The observed association between
overweight/obesity and sciatica did not differ between men
and women and was independent of the designs and response
rates of the included studies, as well as of the assessment
method used for weight and height (self-reported or mea-
sured). Moreover, the associations of overweight/obesity
with lumbar radicular pain and hospitalization due to sciatica
appeared not to be confounded. Exclusion of the studies that
did not control for potential confounders attenuated the asso-
ciation of obesity with clinically defined sciatica only. How-
ever, the subgroup analysis for clinically defined sciatica had
low statistical power because only 2 studies controlled their
estimates for potential confounders.
In conclusion, the findings of this study consistently show
that both overweight and obesity are risk factors for lumbar
radicular pain and sciatica, with a dose-response relationship
among both men and women.
ACKNOWLEDGMENTS
Author affiliations: Centre of Expertise for Health and
Work Ability, Finnish Institute of Occupational Health, Hel-
sinki, Finland (Rahman Shiri, Tea Lallukka, Jaro Karppinen);
Disability Prevention Centre, Finnish Institute of Occupational
Health, Helsinki, Finland (Rahman Shiri, Tea Lallukka, Eira
Viikari-Juntura); and Medical Research Center Oulu, Oulu
University Hospital and University of Oulu, Oulu, Finland
(Jaro Karppinen).
We thank Drs. Bengt Järvholm (16) and Markus Paananen
(22) for providing us with new results.
Conflict of interest: none declared.
REFERENCES
1. Vos T, Flaxman AD, Naghavi M, et al. Years lived with
disability (YLDs) for 1160 sequelae of 289 diseases
and injuries 1990–2010: a systematic analysis for the Global
Burden of Disease Study 2010. Lancet. 2012;380(9859):
2163–2196.
2. Konstantinou K, Dunn KM. Sciatica: review of
epidemiological studies and prevalence estimates. Spine
(Phila Pa 1976). 2008;33(22):2464–2472.
0
0.2
0.4
1.0
0.8
0.6
Standard Error of Log Odds Ratio
–3.0 –2.0 –1.0 0 1.0 2.0 3.0
Log Odds Ratio
Figure 4. Publication bias in studies on the associations of over-
weight/obesity (7 studies) or obesity (19 studies) with lumbar radicular
pain or sciatica, 1966–2013.
Obesity and Sciatica 935
Am J Epidemiol. 2014;179(8):929–937
by guest on March 26, 2015http://aje.oxfordjournals.org/Downloaded from
3. Valat JP, Genevay S, Marty M, et al. Sciatica. Best Pract Res
Clin Rheumatol. 2010;24(2):241–252.
4. Mixter WJ, Barr JS. Rupture of the intervertebral disc with
involvement of the spinal canal. N Engl J Med. 1934;211(5):
210–215.
5. Haldeman S, Kopansky-Giles D, Hurwitz EL, et al.
Advancements in the management of spine disorders. Best
Pract Res Clin Rheumatol. 2012;26(2):263–280.
6. Lean ME, Han TS, Seidell JC. Impairment of health and quality
of life using new US federal guidelines for the identification of
obesity. Arch Intern Med. 1999;159(8):837–843.
7. Kääriä S, Leino-Arjas P, Rahkonen O, et al. Risk factors of
sciatic pain: a prospective study among middle-aged
employees. Eur J Pain. 2011;15(6):584–590.
8. Shiri R, Solovieva S, Husgafvel-Pursiainen K, et al. The role of
obesity and physical activity in non-specific and radiating low
back pain: the Young Finns Study. Semin Arthritis Rheum.
2013;42(6):640–650.
9. Heliövaara M, Mäkelä M, Knekt P, et al. Determinants of
sciatica and low-back pain. Spine (Phila Pa 1976). 1991;16(6):
608–614.
10. Leino-Arjas P, Kauppila L, Kaila-Kangas L, et al. Serum lipids
in relation to sciatica among Finns. Atherosclerosis. 2008;
197(1):43–49.
11. Younes M, Béjia I, Aguir Z, et al. Prevalence and risk factors of
disk-related sciatica in an urban population in Tunisia. Joint
Bone Spine. 2006;73(5):538–542.
12. Shiri R, Solovieva S, Husgafvel-Pursiainen K, et al. Incidence
of nonspecific and radiating low back pain: followup of 24–
39-year-old adults of the Young Finns Study. Arthritis Care Res
(Hoboken). 2010;62(4):455–459.
13. Lequin MB, Verbaan D, Jacobs WC, et al. Surgery versus
prolonged conservative treatment for sciatica: 5-year results of a
randomised controlled trial. BMJ Open. 2013;3(5):e002534.
14. Tubach F, Beauté J, Leclerc A. Natural history and prognostic
indicators of sciatica. J Clin Epidemiol. 2004;57(2):174–179.
15. Kaila-Kangas L, Leino-Arjas P, Karppinen J, et al. History of
physical work exposures and clinically diagnosed sciatica
among working and nonworking Finns aged 30 to 64. Spine
(Phila Pa 1976). 2009;34(9):964–969.
16. Wahlström J, Burström L, Nilsson T, et al. Risk factors for
hospitalization due to lumbar disc disease. Spine (Phila Pa
1976). 2012;37(15):1334–1339.
17. Heliövaara M. Body height, obesity, and risk of herniated
lumbar intervertebral disc. Spine (Phila Pa 1976). 1987;12(5):
469–472.
18. Shiri R, Karppinen J, Leino-Arjas P, et al. Cardiovascular and
lifestyle risk factors in lumbar radicular pain or clinically
defined sciatica: a systematic review. Eur Spine J. 2007;16(12):
2043–2054.
19. Shiri R, Karppinen J, Leino-Arjas P, et al. The association
between obesity and low back pain: a meta-analysis. Am J
Epidemiol. 2010;171(2):135–154.
20. Shiri R, Karppinen J, Leino-Arjas P, et al. The association
between smoking and low back pain: a meta-analysis. Am J
Med. 2010;123(1):87.e7–87.e35.
21. Longo UG, Denaro L, Spiezia F, et al. Symptomatic disc
herniation and serum lipid levels. Eur Spine J. 2011;20(10):
1658–1662.
22. Rivinoja AE, Paananen MV, Taimela SP, et al. Sports,
smoking, and overweight during adolescence as predictors of
sciatica in adulthood: a 28-year follow-up study of a birth
cohort. Am J Epidemiol. 2011;173(8):890–897.
23. Leino-Arjas P, Kaila-Kangas L, Kauppinen T, et al.
Occupational exposures and inpatient hospital care for lumbar
intervertebral disc disorders among Finns. Am J Ind Med. 2004;
46(5):513–520.
24. Saftic´ R, Grgic´ M, Ebling B, et al. Case-control study of risk
factors for lumbar intervertebral disc herniation in Croatian
island populations. Croat Med J. 2006;47(4):593–600.
25. Karjalainen U, Paananen M, Okuloff A, et al. Role of
environmental factors and history of low back pain in sciatica
symptoms among Finnish adolescents. Spine (Phila Pa 1976).
2013;38(13):1105–1111.
26. Armijo-Olivo S, Stiles CR, Hagen NA, et al. Assessment of
study quality for systematic reviews: a comparison of the
Cochrane Collaboration Risk of Bias Tool and the Effective
Public Health Practice Project Quality Assessment Tool:
methodological research. J Eval Clin Pract. 2012;18(1):12–18.
27. Jüni P, Altman DG, Egger M. Assessing the quality of
randomised controlled trials. In: Egger M, Smith GD,
Altman DG, eds. Systematic Reviews in Health Care:
Meta-Analysis in Context. 2nd ed. London, United Kingdom:
BMJ Publishing Group; 2008:87–108.
28. World Health Organization. Physical Status: The Use and
Interpretation of Anthropometry. Report of a WHO Expert
Committee. (WHO Technical Report Series no. 854). Geneva,
Switzerland: World Health Organization; 1995.
29. WHO Expert Consultation. Appropriate body-mass index for
Asian populations and its implications for policy and
intervention strategies. Lancet. 2004;363(9403):157–163.
30. Riihimäki H, Wickström G, Hänninen K, et al. Predictors of
sciatic pain among concrete reinforcement workers and house
painters—afive-year follow-up. Scand J Work Environ Health.
1989;15(6):415–423.
31. Schumann B, Bolm-Audorff U, Bergmann A, et al. Lifestyle
factors and lumbar disc disease: results of a German
multi-center case-control study (EPILIFT). Arthritis Res Ther.
2010;12(5):R193.
32. Kaila-Kangas L, Leino-Arjas P, Riihimäki H, et al.
Smoking and overweight as predictors of hospitalization
for back disorders. Spine (Phila Pa 1976). 2003;28(16):
1860–1868.
33. Vessey M, Painter R, Mant J. Oral contraception and
other factors in relation to back disorders in women:
findings in a large cohort study. Contraception. 1999;60(6):
331–335.
34. Miranda H, Viikari-Juntura E, Martikainen R, et al. Physical
exercise and musculoskeletal pain among forest industry
workers. Scand J Med Sci Sports. 2001;11(4):239–246.
35. Jhawar BS, Fuchs CS, Colditz GA, et al. Cardiovascular risk
factors for physician-diagnosed lumbar disc herniation. Spine J.
2006;6(6):684–691.
36. Cole TJ, Bellizzi MC, Flegal KM, et al. Establishing a standard
definition for child overweight and obesity worldwide:
international survey. BMJ. 2000;320(7244):1240–1243.
37. Manninen P, Riihimäki H, Heliövaara M. Incidence and risk
factors of low-back pain in middle-aged farmers. Occup Med
(Lond). 1995;45(3):141–146.
38. Lipsey MW, Wilson DB. Practical Meta-Analysis. Thousand
Oaks, CA: Sage Publications; 2001.
39. Videman T, Sarna S, Battié MC, et al. The long-term effects of
physical loading and exercise lifestyles on back-related
symptoms, disability, and spinal pathology among men. Spine
(Phila Pa 1976). 1995;20(6):699–709.
40. Higgins J, Green S. Cochrane Handbook for Systematic
Reviews of Interventions. Version 5.1.0 [updated March
2011]. Oxford, United Kingdom: The Cochrane Collaboration;
2009. (www.cochrane-handbook.org/). (Accessed October
29, 2013).
936 Shiri et al.
Am J Epidemiol. 2014;179(8):929–937
by guest on March 26, 2015http://aje.oxfordjournals.org/Downloaded from
41. Mattila VM, Saarni L, Parkkari J, et al. Early risk factors for
lumbar discectomy: an 11-year follow-up of 57,408
adolescents. Eur Spine J. 2008;17(10):1317–1323.
42. Higgins JP, Thompson SG. Quantifying heterogeneity in a
meta-analysis. Stat Med. 2002;21(11):1539–1558.
43. Ioannidis JP, Patsopoulos NA, Evangelou E. Uncertainty in
heterogeneity estimates in meta-analyses. BMJ. 2007;
335(7626):914–916.
44. Morton SC, Adams JL, Suttorp MJ, et al. Meta-Regression
Approaches: What, Why, When, and How? (AHRQ Technical
Reviews, no. 8). Rockville, MD: Agency for Healthcare
Research and Quality, US Department of Health and Human
Services; 2004.
45. Rothstein H, Sutton A, Borenstein M. Publication Bias in
Meta-Analysis: Prevention, Assessment and Adjustments.
Chichester, United Kingdom: John Wiley & Sons Ltd; 2005.
46. Duval S, Tweedie R. Trim and fill: a simple funnel-plot-based
method of testing and adjusting for publication bias in
meta-analysis. Biometrics. 2000;56(2):455–463.
47. Borenstein M, Hedges L, Higgins J, et al. Introduction to
Meta-Analysis. New York, NY: John Wiley & Sons, Inc.; 2009.
48. Leino-Arjas P, Solovieva S, Kirjonen J, et al. Cardiovascular
risk factors and low-back pain in a long-term follow-up of
industrial employees. Scand J Work Environ Health. 2006;
32(1):12–19.
49. Kostova V, Koleva M. Back disorders (low back pain,
cervicobrachial and lumbosacral radicular syndromes)
and some related risk factors. J Neurol Sci. 2001;192(1-2):
17–25.
50. Toda Y, Segal N, Toda T, et al. Lean body mass and body fat
distribution in participants with chronic low back pain. Arch
Intern Med. 2000;160(21):3265–3269.
51. Zhang YG, Sun Z, Zhang Z, et al. Risk factors for lumbar
intervertebral disc herniation in Chinese population: a
case-control study. Spine (Phila Pa 1976). 2009;34(25):
E918–E922.
52. Böstman OM. Body mass index and height in patients requiring
surgery for lumbar intervertebral disc herniation. Spine (Phila
Pa 1976). 1993;18(7):851–854.
53. Hicks GE, Gaines JM, Shardell M, et al. Associations of back
and leg pain with health status and functional capacity of older
adults: findings from the Retirement Community Back Pain
Study. Arthritis Rheum. 2008;59(9):1306–1313.
54. Peul WC, Brand R, Thomeer RT, et al. Influence of gender and
other prognostic factors on outcome of sciatica. Pain. 2008;
138(1):180–191.
55. Zitting P, Rantakallio P, Vanharanta H. Cumulative incidence
of lumbar disc diseases leading to hospitalization up to the age
of 28 years. Spine (Phila Pa 1976). 1998;23(21):2337–2343.
56. Berg AH, Scherer PE. Adipose tissue, inflammation, and
cardiovascular disease. Circ Res. 2005;96(9):939–949.
57. Dumond H, Presle N, Terlain B, et al. Evidence for a key role of
leptin in osteoarthritis. Arthritis Rheum. 2003;48(11):
3118–3129.
58. Li Z, Shen J, Wu WK, et al. The role of leptin on the
organization and expression of cytoskeleton elements in
nucleus pulposus cells. J Orthop Res. 2013;31(6):847–857.
59. Rihn JA, Kurd M, Hilibrand AS, et al. The influence of obesity
on the outcome of treatment of lumbar disc herniation: analysis
of the Spine Patient Outcomes Research Trial (SPORT). J Bone
Joint Surg Am. 2013;95(1):1–8.
60. Meredith DS, Huang RC, Nguyen J, et al. Obesity increases the
risk of recurrent herniated nucleus pulposus after lumbar
microdiscectomy. Spine J. 2010;10(7):575–580.
61. Karppinen J, Kurunlahti M, Taimela S, et al. Determinants of
lumbar artery occlusion among patients with sciatica: a
three-year follow-up with magnetic resonance angiography.
Eur Spine J. 2005;14(7):664–670.
Obesity and Sciatica 937
Am J Epidemiol. 2014;179(8):929–937
by guest on March 26, 2015http://aje.oxfordjournals.org/Downloaded from
