Spine J. 2010 (Oct); 10 (10): 91840
Simon Dagenais DC, PhD, Ralph E. Gay DC, MD, Andrea C. Tricco PhD,
Michael D. Freeman PhD, MPH, DC and John M. Mayer DC, PhD
2732 Transit Rd, West Seneca, NY 14224, USA.
This issue of The Spine Journal includes a review of the scientific evidence supporting spinal manipulative therapy (SMT) for low back pain. The results were quite favorable and reflect a growing body of evidence supporting SMT over medications and other conservative options. SMT research demonstrates equivalent or superior improvement in pain and function when compared with other commonly used interventions, such as physical modalities, medication, education, or exercise, for short, intermediate, and long-term follow-up. The authors conclude by recommending that other health care providers consider SMT as a viable option if self care or education fails to provide pain relief.
Another review of national and international guidelines, published Dec 2010 by Koes et. al. pointed out the disparities between guidelines with respect to spinal manipulation and the use of drugs for both chronic and acute low back pain.
Another review of guidelines published in June 2010 also noted a great degree of similarity between guidelines and that Recommendations for management of acute LBP emphasized patient education, with short-term use of acetaminophen, nonsteroidal anti-inflammatory drugs, or spinal manipulation therapy.
Although there is always a need for more evidence, the evidence over the last few years is providing much stronger support for SMT and that evidence is slowly finding its way into major clinical guidelines both in the United States and internationally.
BACKGROUND CONTEXT: Low back pain (LBP) continues to be a very prevalent, disabling, and costly spinal disorder. Numerous interventions are routinely used for symptoms of acute LBP. One of the most common approaches is spinal manipulation therapy (SMT).
PURPOSE: To assess the current scientific literature related to SMT for acute LBP.
PATIENT SAMPLE: Not applicable.
OUTCOME MEASURES: Not applicable.
DESIGN: Systematic review (SR).
METHODS: Literature was identified by searching MEDLINE using indexed and free text terms. Studies were included if they were randomized controlled trials (RCTs) published in English, and SMT was administered to a group of patients with LBP of less than 3 months. RCTs included in two previous SRs were also screened, as were reference lists of included studies. Combined search results were screened for relevance by two reviewers. Data related to methods, risk of bias, harms, and results were abstracted independently by two reviewers.
RESULTS: The MEDLINE search returned 699 studies, of which six were included; an additional eight studies were identified from two previous SRs. There were 2,027 participants in the 14 included RCTs, which combined SMT with education (n=5), mobilization (MOB) (n=4), exercise (n=3), modalities (n=3), or medication (n=2). The groups that received SMT were most commonly compared with those receiving physical modalities (n=7), education (n=6), medication (n=5), exercise (n=5), MOB (n=3), or sham SMT (n=2). The most common providers of SMT were chiropractors (n=5) and physical therapists (n=5). Most studies (n=6) administered 5 to 10 sessions of SMT over 2 to 4 weeks for acute LBP. Outcomes measured included pain (n=10), function (n=10), health-care utilization (n=6), and global effect (n=5). Studies had a follow-up of less than 1 month (n=7), 3 months (n=1), 6 months (n=3), 1 year (n=2), or 2 years (n=1). When compared with various control groups, results for improvement in pain in the SMT groups were superior in three RCTs and equivalent in three RCTs in the short term, equivalent in four RCTs in the intermediate term, and equivalent in two RCTs in the long term. For improvement in function, results from the SMT groups were superior in one RCT and equivalent in four RCTs in the short term, superior in one RCT and equivalent in one RCT in the intermediate term, and equivalent in one RCT and inferior in one RCT in the long term. No harms related to SMT were reported in these RCTs.
CONCLUSIONS: Several RCTs have been conducted to assess the efficacy of SMT for acute LBP using various methods. Results from most studies suggest that 5 to 10 sessions of SMT administered over 2 to 4 weeks achieve equivalent or superior improvement in pain and function when compared with other commonly used interventions, such as physical modalities, medication, education, or exercise, for short, intermediate, and long-term follow-up. Spine care clinicians should discuss the role of SMT as a treatment option for patients with acute LBP who do not find adequate symptomatic relief with self-care and education alone.
From the FULL TEXT Article:
Low back pain (LBP) is a common and often disabling
condition. The cumulative 1-year incidence of LBP is approximately
20% [1, 2], with most initial episodes being
mild . The reported prevalence of LBP varies greatly.
The point prevalence ranges from 6% to 33% [3, 4] and the
1-year prevalence from 22% to 65% . The lifetime prevalence
of LBP is even more variable, likely because of differences
in the definitions of LBP used, the populations
studied, and the study methodology . There has been a recent
effort to promote a common definition of LBP that will
allow comparisons to be made between studies .
Low back pain is commonly classified as acute (<3
months) or chronic (>3 months) based on its duration .
These temporal definitions appear to be based on studies that
showed that almost all persons with LBP returned to work
within 90 days [8, 9]. Although acute LBP does tend to improve
with time and generally has a good prognosis, improvement
in pain and disability does not correlate well
with return-to-work rates . Furthermore, recent studies
have shown that a significant proportion of acute LBP sufferers
will develop recurrent or chronic LBP. A survey of
persons 35 to 45 years old found that LBP resolved quickly
in only 27% of subjects, whereas 40% developed persistent
LBP . Even among those whose LBP had initially resolved,
29% had recurrent (usually mild) LBP within 6
months . Other studies have found similar trends for recurrence
of LBP [2, 11]. Although it is difficult to predict
who among those with first episodes of LBP will develop recurrent
or chronic symptoms, factors related to the determinants
of disability and to the prediction of chronic disability
appear by 14 days after the onset of pain, supporting that as
a cutoff point in the transition from acute to subacute pain
. Psychological factors appear to play an important role
in that transition and related disability .
Low back pain is a significant societal burden. Persons
seeking care for LBP constitute a substantial proportion
of patients seen in primary care offices. Direct and indirect
costs for LBP have been reported in studies from many
countries, but differences in methodology make it difficult
to compare the results. A recent review suggested that, although
the total yearly cost of LBP (direct and indirect
costs) in the United States has been reported to be between
$19.6 and $118.8 billion per year, the true cost may be
much higher .
Much can be learned from a brief but thorough history
and examination of patients with LBP. Clinical practice
guidelines from the United States and various European
groups suggest that, in the absence of any red flags for
serious spinal pathology, advanced diagnostic studies are
not needed in the initial evaluation of acute LBP [15, 16].
Red flags for LBP are symptoms, findings, or other characteristics
that may be indicative of rare but potentially serious
spinal pathology, such as spinal tumor, infection,
fracture, or cord compromise . Examples of red flags
include unexplained weight loss, loss of bowel or bladder
function, saddle anesthesia, widespread neurologic symptoms
in the lower extremities, recent trauma with osteoporosis
or prolonged corticosteroid use, immune suppression,
and systemic unwellness . Such an evaluation should be
based on the symptoms of the patient and the diagnostic
concerns of the physician but may include X-ray; advanced
imaging (bone scan, computed tomography, or magnetic
resonance imaging); laboratory studies; or electrophysiological
In most cases of acute LBP, an objective cause cannot be
found. Such cases are, therefore, described as nonspecific.
Despite this lack of knowledge regarding the etiology
of LBP, there are many interventions available, and
many providers who are willing to use them . The providers
training often biases the choice of treatment for
acute LBP. Common primary care approaches include education,
reassurance, return to activities, nonsteroidal antiinflammatory
drugs (NSAIDs), and simple analgesics. Patients with acute LBP who do not improve quickly often
seek additional care from both surgical and nonsurgical
specialists. One of the most common treatments used in
North America and Europe is spinal manipulation therapy
(SMT) . Practitioners have used some form of SMT
to treat LBP for thousands of years .
In North America, SMT is usually provided by Doctors of
Chiropractic (DCs) . However, in other countries, particularly
in Europe and Australia, it is commonly used by physical
therapists (PTs), Doctors of Osteopathy (DOs), and
medical doctors (MDs) trained in manual therapy. How
SMT works is not completely understood, but there is growing
evidence that its effects result from a combination of mechanical,
neurological, and biochemical changes in various
structures . Like many therapies administered for acute
LBP, SMT has a diminishing effect size as the duration of
follow-up increases. As a result, its clinical efficacy for acute
LBP is still debated despite many randomized controlled trials
(RCTs), systematic reviews (SRs), and meta-analyses.
The North American Spine Society (NASS) Contemporary
Concepts are a series of evidence-based reviews of
contemporary issues in spine care, intended to provide
spine clinicians with a general understanding about current
practices. Because of the uncertainty of the role of SMT in
the care of acute LBP within the community of spine care
providers at large, a Complementary Medicine Task Force
composed of NASS members (primarily members of the
former NASS Complementary Medicine Committee, see
Acknowledgments) was appointed to develop a Contemporary
Concepts article on SMT for acute LBP.
The eligibility criteria were based on the Population, Intervention,
Control, Outcomes, Study design (PICOS) principle
 as follows:
adults with acute LBP (ie, pain lasting <12
SMT or mobilization (MOB);
any control group that did not receive SMT or
MOB or allowed for evaluation of the comparative efficacy
of different forms of SMT or MOB;
patient-reported pain reduction and functional
improvement (primary outcomes), as well as
global effect, health-care utilization, and harms (secondary
limited to RCTs
Only RCTs published in English were eligible for inclusion.
RCTs were excluded if most of the participants had
symptoms for more than 12 weeks, the study design did
not permit isolating the effects of SMT (or MOB), SMT
was nonforce, participants had multiple indications and
only combined results were reported, studies had fewer
than 20 participants enrolled in each study group, or
follow-up was less than 1 week after the last intervention.
Randomized controlled trials were identified through
an electronic search of MEDLINE (OVID Interface) on
January 13, 2009, and screening the reference lists of two
previous SRs on this topic [23, 24].
MEDLINE was searched using strategies that were validated
by the Cochrane Back Review Group (CBRG) to
identify RCTs (part A of the search strategy) related to spinal
disorders (parts B and D), which were combined with
search terms related to SMT (Table 1) . The following
limits were applied to the search results in OVID: 1) articles
published in English; 2) related to adult participants;
3) published from 1999 to 2009; and 4) abstracts available
Two reviewers independently screened the search results
using predefined inclusion and exclusion criteria. Reviewers
discussed disagreements until consensus was
reached. Full-text articles were retrieved for all citations
deemed relevant or of uncertain relevance to confirm their
eligibility. Reasons for excluding the retrieved full-text articles
Data collection process
One reviewer used a predefined data extraction instrument
to extract the data from the included RCTs. Another
reviewer subsequently verified the data to ensure accuracy.
The main categories of data abstracted for included
studies were methods, pain outcome results, functional
outcome results, other outcome results (eg, global effect,
health-care utilization), and harms.
For methods, the following items were abstracted:
1) study setting and population;
2) inclusion criteria;
3) exclusion criteria;
4) description of intervention(s), provider, and regimen (eg,
frequency, duration) received by experimental group;
5) description of intervention(s), provider, and regimen (eg,
frequency, duration) received by control group(s);
6) number of participants enrolled in experimental group;
7) number of participants enrolled in control group(s);
8) outcome measures and follow-up points.
For pain, functional, and other outcome results, the following data were abstracted:
1) group means at baseline and each follow-up point and
2) statistical significance of intra- and intergroup differences
at baseline and each follow-up point.
For harms, the following data were abstracted:
1) type of adverse event,
2) number of adverse events occurring in the intervention
3) number of adverse events occurring in the
Risk of bias
To assess the possibility of bias in the results of the included
RCTs, methodological quality was assessed according
to a tool developed by the CBRG . The tool includes
12 criteria that assess different aspects of bias, and each
must be scored as yes/no/unsure; RCTs scoring 6 or higher
and without serious flaws were deemed to be of higher
methodological quality . Two reviewers assessed the risk
of bias in the included studies independently. Conflicts between
reviewers were discussed until consensus was
achieved. The clinical relevance of RCTs was also assessed
descriptively according to whether the study protocols reflected
common clinical practice for the application of
SMT for acute LBP.
Synthesis of results
Results related to pain, function, global effect, healthcare
utilization, and other outcomes were synthesized separately
according to whether the experimental group that
received SMT (or MOB) was superior, equal, or inferior
to the control group, as determined by between-group statistical
significance (p<.05) reported in the studies at each
follow-up point. Results were summarized by duration of
follow-up, with less than 4 weeks considered short term,
1 to 6 months considered intermediate term, and more than
6 months considered long term. Pain and functional outcomes
were converted to a 0 to 100 scale, and percent improvement
from baseline was calculated for each follow-up
point. If multiple studies reported results for pain or function
at similar follow-up points within each category of duration
(eg, 12, 34 weeks), the mean and standard
deviation (SD) for improvement from baseline were calculated
for the experimental groups and control groups. If
a study reported results for multiple, similar follow-ups (eg,
1 and 2 weeks), only results for the longest duration within
that timeframe were included in the synthesis. If studies reported
multiple outcome measures for pain (eg, visual analog
scale [VAS] and McGill Pain Questionnaire) or function
(eg, Roland-Morris Disability Questionnaire [RMDQ] and
Oswestry Disability Index [ODI]), results from all outcome
measures reported were included in the synthesis. Results
for outcomes other than pain or function were synthesized
Risk of bias across studies
The risk of bias across studies was assessed using the
outcome-reporting bias item of the CBRG tool.
The MEDLINE search resulted in 699 citations, of
which eight were deemed potentially relevant ,
and 11 were of uncertain relevance . After screening
full-text articles for those 19 studies, only six were
deemed eligible [26, 28, 29, 3133]. Reasons for excluding
full-text articles included duplicate reports (n = 8) [3438, 4042], less than 20 participants per study group
(n = 1) , not being able to distinguish separate effects
of multimodal intervention (n = 1) , mixed neck and
back pain (n = 1) , no patient-reported outcomes
(n = 1) , and no acute LBP (n = 1)  (Figure). An additional
eight eligible studies  were identified from
two previous SRs on similar topics [23, 53]. A total of 14
studies were, therefore, included in this review
[26, 28, 29, 3133, 4552]; only one study was published in
a journal not indexed in MEDLINE . These studies
were published between 1974 and 2006, including one in
the 1970s , five in the 1980s [46, 48, 49, 51, 52], three
in the 1990s [26, 45, 50], and five in the 2000s [28, 29, 3133]. These studies originated in five countries, including
Australia [29, 46], Canada , Italy [33, 52], United Kingdom
[32, 47, 50, 51], and the United States [26, 28, 31, 45, 49]. Study
participants were recruited from primary care settings
(n = 6), outpatient medical centers (n = 3), PT centers (n = 2),
employees of a company (n = 1), or unclear settings (n = 2).
The full study descriptions can be found in Table 2.
Acute LBP was defined using both a minimum duration
(n = 4), ranging from 1 to 4 weeks, and a maximum duration
(n = 5), ranging from 3 weeks to 6 months. All studies enrolled
participants with less than 12 weeks of symptoms,
consistent with parameters set by the CBRG . Only
one study specified that participants were required to have
LBP amenable to SMT to enroll . The number of
exclusion criteria reported ranged from 2 to 15, and the
most common was severe disease (n = 13), followed by
nerve root compression/neurological symptoms (n = 11),
pregnancy (n = 8), prior lumbar surgery (n = 6), systemic inflammation
(n = 5), active litigation/workers compensation
(n = 5), current treatment for LBP (n = 5), mild severity
(n = 5), prior SMT (n = 4), and psychological illness (n = 3).
The experimental group received some form of SMT, including
high-velocity or low-amplitude (HVLA; ie, manual
thrusts pushing the spinal joints slightly beyond their passive
range of motion ) (n = 6) [26, 28, 33, 45, 49, 50]; rotational
(ie, an HVLA technique involving rotating the
patients thigh and leg ) (n = 3) [47, 48, 51]; HVLA or
MOB (ie, manual force to the spinal joints not involving
a thrust or pushing them beyond their passive range of motion
) (n = 2) [29, 32]; instrument (n = 1) ; MOB
(n = 1) ; or unspecified (n = 1) . There were two experimental
groups that received the intervention of interest
(eg, SMT) in two of the included studies [29, 32]. The number
of study treatments in the experimental groups ranged
from 1 to 20 sessions delivered over 1 to 12 weeks, with
most studies providing 5 to 10 sessions [26, 28, 31, 32, 4547, 49] over 2 to 4 weeks [28, 29, 31, 33, 45, 46, 51, 52]. A
DC (n = 5) or PT (n = 5) most frequently administered
SMT, although some studies used a DO (n = 2) or MD
(n = 2); it was unclear who the provider was in one study
(n = 1). The number of participants enrolled in the
experimental groups at baseline ranged from 21 to 165,
with a mean of 63.1 and an SD (6) of 38.1; overall, a total
of 1,009 participants received SMT across these 14 studies.
Studies most frequently had one control group (n = 10),
though some had two (n = 3) or even four (n = 1) control
groups against which SMT was compared. The most common
intervention given to the control groups was physical
modalities (eg, ultrasound, transcutaneous electrical nerve
stimulation, heat, ice) (n = 7), followed by medication (eg,
NSAIDs, muscle relaxants, acetaminophen) (n = 5); education
(n = 6); exercise (eg, strengthening, aerobic, stretching)
(n = 5); MOB (n = 3); and sham SMT (n = 2). Other interventions
as controls to SMT included lumbar supports,
sham physical modalities (eg, detuned diathermy), placebo
medication, and bed rest. The number of treatment sessions
and duration of treatment in the experimental groups generally
mirrored those in the experimental groups (eg, 510
sessions over 24 weeks). The most common type of provider
in the control groups was PT (n = 8), followed by
MD (n = 5), DC (n = 2), and others (n = 1). The total number
of participants enrolled in control groups was higher than in
experimental groups at 1,018, ranging from 23 to 133, with
a mean of 53.6±31.6.
The methodological quality results can be found in
Table 3. The number of criteria met by RCTs varied from
3 to 11, with a mean of 6.6±2.4. Nine RCTs met six or
more items [26, 28, 29, 3133, 45, 47, 49], but two had high
dropouts [31, 32]. There were seven RCTs of higher
methodological quality that met 8.1±2.0 criteria
[26, 28, 29, 33, 45, 47, 49] and seven RCTs of lower methodological
quality that met 5.161.8 criteria [31, 32, 46, 48, 5052]. The quality criterion most commonly fulfilled was
no selective outcome reporting (n = 14), followed by groups
similar at baseline (n = 13), having similar timing of outcome
assessment (n = 12), blinding of outcome assessor
(n = 10), acceptable dropout (n = 10), adequate randomization
(n = 7), intention-to-treat analysis (n = 7), treatment allocation
concealment (n = 5), patient blinded (n = 5), similar
cointerventions (n = 5), and acceptable compliance (n = 5).
The results for the pain outcomes can be found in
Table 4. Six RCTs reported pain outcomes using the
VAS [26, 3133] or numerical rating scale [45, 46]; one
study reported pain relief (%) without raw data . Baseline
pain scores in the experimental groups had a mean of
51.8±4.1; baseline scores were similar in the control
groups (48.4±6.6). Five studies [31, 33, 4547] reported
pain reduction after 1 to 2 weeks of 38±13% in the experimental
groups and 34±19% in control groups; this
difference was statistically significant in two studies
[31, 33]. Four studies [31, 33, 45, 46] reported pain reduction
after 3 to 4 weeks of 66±17% in the experimental groups
and 50±22% in control groups; this difference was statistically
significant in one study . Four studies
[26, 32, 33, 45] reported pain reduction after 2 to 3 months
of 57±13% in the experimental groups and 46±10% in
control groups; this difference was statistically significant
in one study .
Two studies [32, 33] reported pain reduction after 6
months of 47±19% in the experimental groups and
46±1% in control groups; these differences were not statistically
significant. Two studies [32, 45] reported pain reduction
after 1 year of 51±16% in the experimental groups
and 55±9% in control groups; these differences were not
statistically significant. One study  reported pain reduction
after 2 years of 71±0.0% in the experimental group
and 60±10% in the control groups; this difference was
not statistically significant. Other outcomes related to pain
included LBP recurrence [32, 45], which was 63.5±19.1%
in experimental groups and 58.3±9.7% in control groups
after 1 year and 70±0.0% in both groups after 2 years,
and participants recovered from symptoms [29, 33, 51],
whose values were 40.5±31.5% in the experimental groups
and 37.8±31.3% in the control groups after 2 to 4 weeks;
26.0±2.8% and 6.0±0.0%, respectively, after 3 to 6
months (this difference was statistically significant); and
46.5±2.1% and 32.5±6.4%, respectively, after 1 year.
The functional outcome results can be found in Table 5.
Nine RCTs reported functional outcomes using the RMDQ
[26, 29, 32, 45, 49], ODI [26, 28, 31], Short-Form 36 (SF-36)
physical function [32, 33], or other outcome measures
[29, 50]; one reported only the statistical significance of differences
without raw data . Some studies reported multiple
functional outcomes [26, 29, 32], and others reported
functional outcome results for multiple subgroups
[49, 50]. Baseline RMDQ scores from five experimental
groups had a mean of 43.3 and an SD of 8.1; baseline
scores were similar in seven control groups
(mean±SD = 43.5±8.6). Baseline ODI scores from three
experimental groups had a mean of 60.7±18.5; scores were
similar in four control groups (56.0±17.4). Five studies
[28, 31, 45, 49, 50] reported functional improvement in the
experimental groups of 53±15% after 1 to 2 weeks and
44±22% in the control groups; this difference was statistically
significant in one study . Four studies
[28, 31, 45, 50] reported functional improvement after 3 to
4 weeks of 55±7% in the experimental groups and
55±21% in control groups; this difference was statistically
significant in one study . Three studies [26, 32, 45] reported
functional improvement after 2 to 3 months of
51±21% in the experimental groups and 58±23% in the
control groups; these differences were not statistically
Two studies [28, 32] reported functional improvement after
6 months of 44±22% in the experimental groups and
38621% in the control groups; this difference was statistically
significant in one study . Two studies [32, 45] reported
functional improvement after 1 year of 43±23% in
the experimental groups and 49±25% in the control
groups; this difference was statistically significant in one
study . One study  reported functional improvement
after 2 years of 7560% in the experimental group
and 69±9% in the control groups; this difference was not
statistically significant. Other functional outcomes reported
included need for bed rest , which was 8.0±0% in the
experimental group and 10.0±1.4% in control groups after
1 year; reduced activity because of LBP , which was
33±0% in the experimental group and 36±1% in control
groups after 1 year; marked or moderate improvement in
a functional scale of 10 activities of daily living , which
was 29% in the experimental group and 30% in the control
group after 2 weeks.
The global effect results can be found in Table 6. Five
RCTs reported global effect using a variety of scales (eg,
global impression of severity; global rating of care; composite
score combining pain, stiffness, and tenderness)
[29, 31, 45, 48, 52]. Results for global effect were statistically
significant in favor of the experimental groups in two of
five RCTs after 1 to 2 weeks [29, 31, 45, 48], in two of three
RCTs after 3 to 4 weeks [29, 45, 52], and one of two RCTs
after 2 to 3 months [29, 52]; no differences were noted beyond
6 months of follow-up .
The health-care utilization results can be found in Table 7.
Six RCTs reported on health-care utilization for LBP, including
analgesic medication use [28, 3133, 51] and other
health-care use [28, 45]. For analgesic medication use,
there were no statistically significant differences between
groups after 2 weeks in three RCTs [31, 33, 51] nor after
2 months in one RCT . Results after 6 months favored
the experimental group in one RCT , though no differences
were noted in another RCT after 1 year . For
other health-care use, results favored the experimental
group after 6 months in one RCT ; no differences
were noted in one RCT after 1 year .
In addition to pain, function, global effect, and healthcare
utilization, other outcomes were also measured and reported
in these RCTs, and their results can be found in
Table 8. These outcomes included lumbosacral range of
motion (eg, degrees of flexion, extension, and rotation, or
fingertip-to-floor distance) [26, 46, 48]; mental health (eg,
SF-36 mental composite score, modified Zung Questionnaire)
[31,32]; lost work time because of LBP [28, 32, 45];
and utility (eg, EuroQol-5D) . Results for lumbosacral
range of motion were not statistically significant between
groups after 1, 2, 3 weeks, or 3 months [26, 46, 48]. Results
for mental health were no different between groups after 2,
4 weeks or 2 or 6 months  but were statistically significant
in favor of the control group after 1 year . Results
for work loss because of LBP were statistically
significant in favor of the experimental group after 6
months , though no differences were noted between
groups after 1 year [32, 45]. Results for utility were not
statistically significant between groups after 2 or 6 months
or 1 year .
One RCT reported pain outcomes separately for subgroups
that had either a positive or a negative straight
leg raise (SLR) . Results suggest that a greater proportion
had recovered after 2 weeks in those with a positive
SLR, though there were no apparent differences after 1
year . Two RCTs reported functional outcomes separately
for subgroups that had LBP for less than 2 weeks
and for 2 to 4 weeks [49, 50]. Results suggest that a greater
improvement in RMDQ was noted in those with LBP of
2 to 4 weeks duration after 3 days ; no further improvements
were observed in this subgroup after 12 days,
whereas those with LBP duration of less than 2 weeks
noted additional improvement [49, 50]. One RCT reported
global effect outcomes separately for subgroups with or
without leg pain . Results suggest that the experimental
group was more superior to the control groups
in those with leg pain after 3 weeks but not after 2 or
6 months .
Overall results for pain and functional outcomes are
summarized in Tables 9 and 10, respectively.
Data on harms are presented in Table 11. None of the
included studies reported harms specifically related to SMT.
Improvements in pain and function
Spinal manipulation therapy appears to be effective for
pain reduction in the short, intermediate, and long term. Only
1 to 2 weeks after initiating care with SMT, pain reduction
was substantial (62%), though it was almost as large for the
control groups against which it was compared. Pain reduction
tended to peak within 3 to 4 weeks of beginning SMT (80%)
and tapered slightly after 2 to 3 months (67%) and 6 months
(65% SMT) but remained higher than that achieved after 1 to
2 weeks. Pain reduction continued to taper after 1 year (51%)
and 2 years (66%). One-third of the studies that reported pain
outcomes demonstrated a greater pain reduction for SMT
than that for the control groups at one or more time points
[31, 33, 45], whereas two-thirds showed no difference between
SMTand control groups [26, 29, 32, 46, 47, 51]. No studies
reported that SMT was inferior to other treatments in
providing pain reduction at any time point. Of the studies that
reported function and disability outcomes, most (seven out of
nine) reported no difference compared with various control
interventions [26, 31, 33, 45, 4850]. One study demonstrated
a greater improvement for SMT at two time points , and
one study demonstrated a greater improvement in control
compared with SMT at one time point .
Results from most studies indicate thatSMTwas either superior
or equivalent to many commonly used interventions,
including physical modalities, education, exercise, and
medication. In cases where SMT was equivalent to the control
group, both groups improved. Lack of superiority for
any particular approach is likely related to many conservative
interventions having equally favorable results in the initial
stages of LBP. This suggests that a patient with acute LBP
(or a spine clinician involved in their care) can reasonably
choose the most appealing of these management options
based on availability, personal preference, expectation of improvement,
or other factors beyond simply efficacy.
Treatment of acute low back pain
This review is unable to address a different research
question that may also be of interest, that is, is any intervention
at all necessary for acute LBP? The prognosis
of acute LBP is generally viewed as favorable, with or
without treatment. One could, therefore, suggest that the
efficacy of SMT relative to other commonly used interventions
is less important than that compared with no
treatment at all. However, this question can only be addressed
by RCTs in which SMT is compared with a pure
no-treatment control group, which did not occur in the 14
No-treatment control group
Although conducting an RCT of SMT compared with
a no-treatment control group for acute LBP could be of scientific
or economic interest, it may pose certain challenges.
For example, it may be difficult to justify randomizing participants
to a no-treatment control group when several commonly
used interventions, including SMT, appear to be
relatively safe, effective, inexpensive, and widely available.
Even if an ethics review board approved such a study design,
it may be challenging to recruit participants. At minimum,
brief education of participants may be required as to
why a no-treatment approach is appropriate, in effect,
changing the no-treatment control group into a briefeducation
control group. Once recruited, monitoring and
compliance of participants assigned to a no-treatment control
group could also be problematic as they may resort to
self-prescribed nonstudy interventions (eg, over-thecounter
analgesics) for pain relief. This issue may warrant
further discussion among clinicians, researchers, and ethicists,
but cannot be addressed from this review of currently
Previous systematic reviews on this topic
The present review uncovered five new RCTs [28, 29, 3133] that were conducted after publication of two previous
SRs on the same topic [23, 53]. The present review focused
exclusively on acute LBP, whereas the two previous reviews
included acute, subacute, and chronic LBP. Nevertheless,
our results appear to be generally consistent with
those of the two previous SRs. It is important to note, however,
that interpretation of the results and conclusions
drawn from them are different among the SRs. For example,
the present review concludes that SMT is equally effective
as other commonly used conservative approaches for
acute LBP. On the other hand, a previous SR interpreted
similar evidence and concluded that SMT had no statistically
or clinically significant advantage over general medical
care, analgesic medication, physical therapy, exercises,
or back school . Although both conclusions are similar,
their wording may influence how their findings will be perceived
by those who do not read the full study report.
Systematic versus narrative reviews
There are many different ways to approach a review article.
A narrative review simply proposes one or more hypotheses
and cites studies supporting those points. On the
other hand, an SR approaches the process in an organized
and transparent manner that removes much of the bias that
can be introduced in a narrative review. When conducted
appropriately, SRs are considered to be an important tool
in evidence-based medicine. The main steps required in
an SR are to define a research question; devise a comprehensive
and transparent search strategy to uncover relevant
studies; specify study eligibility criteria; screen results independently
by two reviewers to avoid bias in selecting
studies; evaluate methodological quality; summarize results
for similar studies; and interpret findings for those who do
not have the time, expertise, or willingness to do so. Embedded
within each of these steps are decisions that authors
must make that can impact their findings. Readers should
be aware of these potential limitations.
Numerous duplicate reports were uncovered in the
search. Many of these reports were related to the same
cohort of participants but reported new analyses that attempted
to explain observed differences by examining various
hypotheses, including gender , confidence ,
satisfaction , and lumbar mobility . Although secondary
analyses can provide greater insight when interpreting
results, duplicate reports can also artificially inflate the
amount of evidence supporting an intervention if readers
are not aware whether they are referring to the same patient
Many of the RCTs appear to have been designed pragmatically
rather than using standardized methodology. For
example, acute LBP is a fairly simple, universal concept
typically defined as duration of symptoms of less than 12
weeks . However, many RCTs have defined acute LBP
in differing ways that preclude combining their results in
SRs or meta-analyses. When designing future RCTs, investigators
should consider how data from their study might be
combined with data from other studies.
It has been estimated that 94% of SMT is administered
by DCs . However, only 38% (5 out of 13) of the studies
which reported provider type involved DCs as providers
of SMT. Potential explanations for this observation include
insufficient funding of RCTs outside medical research universities,
inadequate research training or infrastructure for
DCs to conduct RCTs, or a perception that efficacy of
SMT for acute LBP has been sufficiently answered and is
no longer of primary research interest. The involvement
of PTs as authors or providers of SMT was greater than that
one might expect, as they were involved in 38% (5 out of
13) of the studies, including three of the most recent RCTs
[28, 29, 32]. It is unknown if results from SMT administered
by different providers are interchangeable in terms of efficacy.
Nonetheless, it is interesting to note that SMT was administered
by DCs in all of the studies that reported greater
pain reduction in the SMT group over control groups at one
or more time points [31, 33, 45].
Frequency and duration of spinal manipulation therapy
For acute LBP, it is common practice to recommend two to
three sessions per week for the first few weeks and then to
gradually decrease the frequency of treatments during subsequent
weeks. In the reviewed studies, the total number of
SMT sessions ranged from 1 [47, 49] to 20 sessions over 30
days . Data on SMT frequency and duration were ambiguous
in some of the reviewed studies; hence, the extent to
which these variables affected outcomes is uncertain. However,
there is no evidence to suggest that 20 treatment sessions
offer clear advantages over 5 to 10 treatment sessions.
Cointerventions, including passive physical modalities,
massage, assisted stretching, exercises, and medications,
were often administered concurrently with SMT. Investigators
who design RCTs that use multiple cointerventions
likely do so to reflect how SMT is used in clinical practice.
However, this practice draws on anecdotal evidence and
other factors as patients attempt to find symptomatic relief.
Administering numerous interventions in RCTs without appropriate
control groups (eg, placebo) obscures the unique
contribution of specific therapies, including those of interest
to investigators (ie, SMT). It may even increase the possibility
that no difference will be found among the various
treatments. The end result is that such studies are typically
not useful in the clinical decision-making process. Monitoring
the use of cointerventions would likely require a participant
diary that could be verified through claim records.
This item is challenging to assess when the control group
intervention is a cointervention in the experimental group
(eg, analgesics). In such cases, control group use should
not be comparable among study arms. Studies with protocols
requiring multiple sessions should report compliance
to help determine if it affects outcome. This would also allow
subgroup analyses of dose and frequency effects.
Spinal manipulation therapy techniques
Many different SMT techniques are used for LBP by
a variety of providers. However, most of the studies we reviewed
did not include sufficient details of the technique
used. One study did directly compare two common techniques
(HVLA SMT and rotational MOB), but only one
treatment session of SMT was carried out in this study
. Therefore, we were not able to draw conclusions
about the relative efficacy of different SMT techniques.
We found that most of the new CBRG criteria were useful
in assessing methodological quality. However, it was difficult
to evaluate the new CBRG criterion regarding selective outcome
reporting, which was assumed to be absent unless outcomes
were specified in the objectives but not reported.
Presumably, the authors intent on concealing unfavorable
results would not mention those outcomes at all rather than
selectively mention in portions of the manuscript that could
create doubts among readers. Furthermore, the relevance of
blinding for physical interventions, such asSMT, is questionable.
Providers cannot readily be blinded to a skilled manual
technique they deliver. Participants cannot easily be blinded
unless a sham SMT is devised, which may itself have therapeutic
effect. When the primary outcomes are selfreported,
blinding the outcome assessor is also less relevant.
Nevertheless, these CBRG criteria are widely used.
The source of funding is an important methodological
consideration because of the potential for funding bias
[58,59]. The source of funding for the included studies is presented
in Table 12. All of the included studies reported a funding
source, which was either a not-for-profit agency (n = 11)
or a government organization (n = 5). Three of the included
studies also reported an in-kind donation or the use of equipment
for theRCTfrom private industry, indicating that the results
of this review are likely not influenced by funding bias.
Spinal manipulation therapy appears to be relatively
safe, as no harms were attributed to SMT in the five studies
that reported harms data. Because RCTs are typically not
powered to estimate the risk of harms, the literature on this
topic was also briefly reviewed to provide additional information
about the known or potential harms of SMT. After
reviewing this literature, it appears that harms associated
with SMT can be divided into relatively common, minor,
temporary, and self-limiting harms (eg, side effects), or
very rare, more serious adverse events (SAEs).
Minor, temporary, and self-limiting harms
The minor, temporary, and self-limiting harms that have
been reported after lumbar SMT include local discomfort,
stiffness, radiating pain, and fatigue; these symptoms are typically
reported to last between several hours and a few days
[8,60]. These minor, temporary, and self-limiting harms have
been reported by 30% to 50% of those receiving lumbar SMT
and have also been reported more frequently in females .
Very rare serious adverse events
In contrast, very rare SAEs have only been reported in
the literature through case reports or case series. The main
types of SAEs associated with lumbar SMT are lumbar disc
injury, cauda equina syndrome, spinal cord ischemia or infarct,
vertebral fracture, and epidural hematoma .
These very rare SAEs are reported so infrequently that
few risk factors have yet been identified, though anti-coagulation
therapy may be associated with epidural hematoma.
In addition, the frequency of very rare SAEs is difficult
to estimate with any precision, as this requires estimating
both a very small numerator (ie, reported SAEs) and a very
large denominator (ie, total number of lumbar SMT in
a given period). One review reported that the rate of disc
herniation or cauda equina syndrome after lumbar SMT
was 1 per 3.7 million procedures; the confidence interval
around this estimate is unknown but likely to be wide .
Based on the RCTs reviewed, SMT appears to be effective
for pain reduction in the short, intermediate, and long
terms. One-third of the studies included in this SR demonstrated
more pain reduction with SMT than for control
groups at one or more time points, whereas two-thirds
showed no difference between SMT and the control groups.
No study found SMT to be inferior to other treatments in
regard to pain reduction at any time. There is no evidence
to suggest that a higher number of treatment sessions with
SMT was superior to the commonly used 5 to 10 treatment
sessions. With the currently available evidence, the choice
of SMT versus other treatment approaches for acute LBP
cannot be made on the basis of relative efficacy alone. That
decision must, therefore, be based on patient preference,
treatment availability, treatment cost, or other factors.
The authors would like to thank all members of the NASS Complementary Medicine Task force for their advice and work on this project (in alphabetical order): Thiru M. Annaswamy, MD; Jay E. Bowen, DO; Simon Dagenais, DC, PhD; Michael D. Freeman, PhD, DC, MPH; Mark R. Foster, MD, PhD; Kim J. Garges, MD, DC; Ralph E. Gay, MD, DC; John M. Mayer, PhD, DC; Steven A. Schopler, MD. The authors would also like to thank their NASS staff liaison, Karen James, for her efforts on this project.
Waxman R, Tennant A, Helliwell P.
A prospective follow-up study of low back pain in the community.
Cassidy JD, Cote P, Carroll LJ, Kristman V.
Incidence and course of low back pain episodes in the general population.
Spine 2005;30: 281723.
Loney PL, Stratford PW.
The prevalence of low back pain in adults: a methodological review of the literature.
Phys Ther 1999;79: 38496.
The prevalence of low back pain: a systematic review of the literature from 1966 to 1998.
J Spinal Disord 2000;13:20517.
Linton SJ, Ryberg M.
Do epidemiological results replicate? The prevalence and health-economic consequences of neck and back pain
in the general population.
Eur J Pain 2000;4:34754.
Dionne CE, Dunn KM, Croft PR, et al.
A consensus approach toward the standardization of back pain definitions for use in prevalence studies.
Furlan AD, Pennick V, Bombardier C, van TM.
2009 updated method guidelines for systematic reviews in the Cochrane Back Review Group.
Andersson GB, Svensson HO, Oden A.
The intensity of work recovery in low back pain.
Atroshi I, Andersson IH, Gummesson C, et al.
Primary care patients with musculoskeletal pain. Value of health-status and sense-ofcoherence measures in predicting long-term work disability.
Scand J Rheumatol 2002;31:23944.
Roland M, Morris R.
A study of the natural history of low-back pain. Part II: development of guidelines for trials of treatment in primary care.
Stanton TR, Henschke N, Maher CG, et al.
After an episode of acute low back pain, recurrence is unpredictable and not as common as previously thought.
Kovacs FM, Abraira V, Zamora J, Fernandez C
Spanish Back Pain Research N. The transition from acute to subacute and chronic low back pain: a study based on determinants of quality of life and prediction of chronic disability.
Manek NJ, MacGregorAJ.
Epidemiology of back disorders: prevalence, risk factors, and prognosis.
Curr Opin Rheumatol 2005;17:13440.
Dagenais S, Caro J, Haldeman S.
A systematic review of low back pain cost of illness studies in the United States and internationally.
Spine J 2008;8:820.
Chou R, Qaseem A, Snow V, et al.
Diagnosis and Treatment of Low Back Pain: A Joint Clinical Practice Guideline
from the American College of Physicians and the American Pain Society
Annals of Internal Medicine 2007 (Oct 2); 147 (7): 478491
van Tulder MW, Becker A, Bekkering T, et al.
European Guidelines for the Management of
Acute Nonspecific Low Back Pain in Primary Care
European Spine Journal 2006 (Mar); 15 Suppl 2: S169191
Dagenais S, Tricco AC, Haldeman S.
Synthesis of Recommendations for the Assessment and Management of
Low Back Pain from Recent Clinical Practice Guidelines
Spine J. 2010 (Jun); 10 (6): 514529
Haldeman S, Dagenais S.
A supermarket approach to the evidence-informed management of chronic low back pain.
Spine J 2008;8:17.
Bronfort G, Haas M, Evans R, et al.
Evidence-informed management of chronic low back pain with spinal manipulation and mobilization.
Spine J 2008;8:21325.
Wiese G, Challender A.
History of spinal manipulation.
In: Haldeman S, Dagenais S, Budgell B, eds.
Principles and practice of chiropractic. 3rd ed.
New York, NY: McGraw-Hill, 2005.
Shekelle PG, Adams AH, Chassin MR, et al.
Spinal manipulation for low-back pain.
Ann Intern Med 1992;117:5908.
Popping the (PICO) question in research and evidencebased practice.
Appl Nurs Res 2002;15:1978.
Assendelft WJ, Morton SC, Yu EI, et al.
Spinal manipulative therapy for low back pain.
Cochrane Database Syst Rev 2004:CD000447.
Bronfort G, Evans R, Maiers M, Anderson AF.
Spinal Manipulation, Epidural Injections, and Self-care
A Pilot Study for a Randomized Clinical Trial
J Manipulative Physiol Ther. 2004 (Oct); 27 (8): 503508
van Tulder M, Furlan A, Bombardier C, Bouter L.
Updated method guidelines for systematic reviews in the Cochrane Collaboration Back Review Group.
Andersson GB, Lucente T, Davis AM, et al.
A comparison of osteopathic spinal manipulation with standard care for patients with low back pain.
N Engl J Med 1999;341:142631.
Bronfort G, Evans RL, Anderson AV, Schellhas KP, Garvey TA, Marks RA, et al.
Nonoperative Treatments for Sciatica: A Pilot Study
for a Randomized Clinical Trial
J Manipulative Physiol Ther. 2000 (Oct); 23 (8): 536544
Childs JD, Fritz JM, Flynn TW, et al.
A Clinical Prediction Rule To Identify Patients
With Low Back Pain Most Likely
To Benefit from Spinal Manipulation: A Validation Study
Annals of Internal Medicine 2004 (Dec 21); 141 (12): 920928
Hancock MJ, Maher CG, Latimer J, et al.
Assessment of diclofenac or spinal manipulative therapy, or both, in addition to recommended first-line treatment for acute low back pain: a randomised controlled trial.
Hay EM, Mullis R, Lewis M, et al.
Comparison of physical treatments versus a brief pain-management programme for back pain in primary care: a randomised clinical trial in physiotherapy practice.
Hoiriis KT, Pfleger B, McDuffie FC, et al.
A Randomized Clinical Trial Comparing Chiropractic Adjustments to Muscle Relaxants
for Subacute Low Back Pain
J Manipulative Physiol Ther 2004 (Jul); 27 (6): 388-398
Hurley DA, McDonough SM, Dempster M, et al.
A randomized clinical trial of manipulative therapy and interferential therapy for acute low back pain.
Santilli V, Beghi E, Finucci S.
Chiropractic manipulation in the treatment of acute back pain and sciatica with disc protrusion: a randomized
double-blind clinical trial of active and simulated spinal manipulations.
Spine J 2006;6:1317.
Flynn TW, Childs JD, Fritz JM.
The audible pop from high-velocity thrust manipulation and outcome in individuals with low back pain.
J Manipulative Physiol Ther 2006;29:405.
Fritz JM, Whitman JM, Childs JD.
Lumbar spine segmental mobility assessment: an examination of validity for determining intervention strategies in patients with low back pain.
Arch Phys Med Rehabil 2005;86:174552.
George SZ, Fritz JM, Childs JD, Brennan GP.
Sex differences in predictors of outcome in selected physical therapy interventions for acute low back pain.
J Orthop Sports Phys Ther 2006;36:35463.
Goldstein MS, Morgenstern H, Hurwitz EL, Yu F.
The impact of treatment confidence on pain and related disability among patients with low-back pain: results from the University of California, Los Angeles, low-back pain study.
Spine J 2002;2:3919; discussion 399401.
Hurwitz EL, Morgenstern H, Chiao C.
Effects of Recreational Physical Activity and Back Exercises
on Low Back Pain
and Psychological Distress: Findings from the UCLA Low Back Pain Study
Am J Public Health. 2005 (Oct); 95 (10): 18171824
Hurwitz EL, Morgenstern H, Vassilaki M, Chiang L-M.
Adverse reactions to chiropractic treatment and their effects on satisfaction and clinical outcomes among patients enrolled in the UCLA Neck Pain Study.
J Manipulative Physiol Ther 2004;27:1625.
Hurwitz EL, Morgenstern H, Yu F.
Satisfaction as a predictor of clinical outcomes among chiropractic and medical patients enrolled in the UCLA Low Back Pain Study.
Kominski GF, Heslin KC, Morgenstern H, et al.
Economic evaluation of four treatments for low-back pain: results from a randomized controlled trial.
Med Care 2005;43:42835.
Seferlis T, Lindholm L, Nemeth G.
Cost-minimisation analysis of three conservative treatment programmes in 180 patients sick-listed for acute low-back pain.
Scand J Prim Health Care 2000;18:537.
Shearar KA, Colloca CJ, White HL.
A Randomized Clinical Trial of Manual Versus Mechanical Force Manipulation
in the Treatment of Sacroiliac Joint Syndrome
J Manipulative Physiol Ther 2005 (Sep);28 (7): 493501
Skillgate E, Vingard E, Alfredsson L.
Naprapathic manual therapy or evidence-based care for back and neck pain: a randomized, controlled trial.
Clin J Pain 2007;23:4319.
Cherkin DC, Deyo RA, Battie M, et al.
A comparison of physical therapy, chiropractic manipulation, and provision of an educational booklet for the treatment of patients with low back pain.
N Engl J Med 1998;339:10219.
Farrell JP, Twomey LT. Acute low back pain.
Comparison of two conservative treatment approaches.
Med J Aust 1982;1:1604.
Glover JR, Morris JG, Khosla T.
Back pain: a randomized clinical trial of rotational manipulation of the trunk.
Br J Ind Med 1974;31: 5964.
Godfrey CM, Morgan PP, Schatzker J.
A randomized trial of manipulation for low-back pain in a medical setting.
Hadler NM, Curtis P, Gillings DB, Stinnett S.
A benefit of spinal manipulation as adjunctive therapy for acute low-back pain: a stratified controlled trial.
MacDonald RS, Bell CM.
An open controlled assessment of osteopathic manipulation in nonspecific low-back pain.
Spine 1990;15: 36470.
Mathews JA, Mills SB, Jenkins VM, et al.
Back pain and sciatica: controlled trials of manipulation, traction, sclerosant and epidural injections.
Br J Rheumatol 1987;26:41623.
Postacchini F, Facchini M, Palieri P.
Efficacy of various forms of conservative treatment in low back pain.
Neuro Orthop 1988;6:2835.
Bronfort G, Haas M, Evans RL, Bouter LM.
Efficacy of Spinal Manipulation and Mobilization
for Low Back Pain and
Neck Pain: A Systematic Review and Best Evidence Synthesis
Spine J (N American Spine Soc) 2004 (May); 4 (3): 335356
Multiple publication of reports of drug trials.
Eur J Clin Pharmacol 1989;36:42932.
Tramer MR, Reynolds DJ, Moore RA, McQuay HJ.
Impact of covert duplicate publication on meta-analysis: a case study.
BMJ 1997;315: 63540.
von Elm E, Poglia G, Walder B, Tramer MR.
Different patterns of duplicate publication: an analysis of articles used in systematic reviews.
Shekelle PG, Adams AH, Chassin MR, et al.
The Appropriateness of Spinal Manipulation for Low Back Pain:
Santa Monica, CA: RAND, 1991. Report No.: R-4025/2-CCR/FCER.
Lexchin J, Bero LA, Djulbegovic B, Clark O.
Pharmaceutical industry sponsorship and research outcome and quality: systematic review.
Bekelman JE, Li Y, Gross CP.
Scope and impact of financial conflicts of interest in biomedical research: a systematic review.
Senstad O, Leboeuf-Yde C, Borchgrevink C.
Frequency and characteristics of side effects of spinal manipulative therapy.
Spine (Phila, PA, 1976) 1997;22:43540.
Barrett AJ, Breen AC.
Adverse effects of spinal manipulation.
J R Soc Med 2000;93:2589.
Gallinaro P, Cartesegna M.
Three cases of lumbar disc rupture and one of cauda equina associated with spinal manipulation (chiropraxis).
Oppenheim JS, Spitzer DE, Segal DH.
Nonvascular complications following spinal manipulation.
Spine J 2005;5:6606.
Dan NG, Saccasan PA.
Serious complications of lumbar spinal manipulation.
Med J Aust 1983;2:6723.
Haldeman S, Rubinstein SM.
Cauda equina syndrome in patients undergoing manipulation of the lumbar spine.
Haldeman S, Rubinstein SM.
Compression fractures in patients undergoing spinal manipulative therapy.
J Manipulative Physiol Ther 1992;15:4504.
Morandi X, Riffaud L, Houedakor J, et al.
Caudal spinal cord ischemia after lumbar vertebral manipulation.
Joint Bone Spine 2004;71: 3347.
Balblanc JC, Pretot C, Ziegler F.
Vascular complication involving the conus medullaris or cauda equina after vertebral manipulation for an L4-L5 disk herniation.
Rev Rhum Engl Ed 1998;65:27982.
Whedon JM, Quebada PB, Roberts DW, Radwan TA.
Spinal epidural hematoma after spinal manipulative therapy in a patient undergoing anticoagulant therapy: a case report.
J Manipulative Physiol Ther 2006;29:5825.
Solheim O, Jorgensen JV, Nygaard OP.
Lumbar epidural hematoma after chiropractic manipulation for lower-back pain: case report.
Safety of spinal manipulation in the treatment of lumbar disk herniations: a systematic review and risk assessment.
J Manipulative Physiol Ther 2004;27:197210.
Return to the LOW BACK PAIN Section