Spine J. 2008 (Jan); 8 (1): 266–277 ~ FULL TEXT
Scott Haldeman, DC, MD, PhD, FRCP(C), Simon Dagenais, DC, PhD
Department of Neurology,
University of California,
Irvine, CA, USA
From the Full-Text Article:
This special focus issue has presented information on 24
categories of treatments that are widely prescribed for the
management of chronic low back pain (CLBP) without surgery, and also provided an overview of commonly available
surgical options. The authors of each of these papers have
spent a great deal of time and effort discussing these treatment approaches and providing their interpretation of the
evidence that is available to justify their use. The fact that
there are 25 categories of treatment presented in this special
focus issue, each of which has multiple subcategories, is
a testament to the fact that no single approach has yet been
able to demonstrate its definitive superiority. This situation
makes it very challenging for clinicians, policy makers, insurers, and patients to make decisions regarding which
treatment is the most appropriate for CLBP.
Although readers may be tempted to examine only those
articles describing their favorite (or least favorite) treatments to find evidence that simply affirms their beliefs, it is highly recommended that the entire special focus issue
be perused to compare and contrast the theories and evidence supporting all approaches. This can help overcome
our natural tendencies to support only those treatments with
which we are most familiar and dismiss those about which
we know little. Only when reasonably informed about all
available treatments will purchasers (eg, patients, insurers)
and providers of care truly understand the current state of
the science and art and be in a position to compare and
make decisions concerning the treatment options for CLBP.
This article will attempt to facilitate this task by summarizing some of the pertinent information from each of the
articles presented in this special focus issue,
Articles from expert clinicians
Readers should be aware of possible biases inherent to
the type of review articles contained in this special focus
issue, many of which were contributed by authors known
to have an interest in specific interventions. These authors
are naturally more likely to be optimistic about the benefits
of a procedure than others who are offering a different
treatment approach. Presumably, clinicians who contributed
articles on specific interventions used in their practice
would not be offering them to their patients if they were
not enthusiastic about their superiority over other options.
This enthusiasm is seen most commonly in articles with extensive discussions about the theoretical basis of a treatment
approach for which there is little available evidence of
efficacy. Similarly, this zest may be at play when authors
attempt to minimize or criticize the importance of clinical
trials that reported negative outcomes they feel are not
reflective of what is observed in their daily practice.
Articles by clinical researchers
A number of these review articles were written by authors
who work primarily as researchers and are not involved in
clinical practice offering the treatments about which they
wrote. These authors may exhibit other biases to their interpretation of the scientific literature supporting or refuting the efficacy of a treatment approach. Whereas clinicians tend to be overly optimistic about the efficacy of an intervention based mainly on their personal experience, researchers (many of whom are clinical epidemiologists) tend to be overly pessimistic about interventions for which there is little, low-quality, or conflicting evidence of efficacy. This is based, in part, on medicine’s long history of once promising
but eventually discredited treatments, and the principle of
primum non nocere. Although the former often discount research evidence, the latter often overlook clinical experience; neither viewpoint is ideal.
Systematic reviews (SRs) conducted on intervention for
CLBP that adhered strictly to the principles and rigorous
methodology of evidence-based medicine (EBM) often
conclude with the statement that there is insufficient evidence and more research is necessary. Although the efforts
of EBM to improve the practice of health care are laudable,
decisions must still be made on a daily basis in the absence
of the amount and quality of evidence necessary to convince clinical epidemiologists that an intervention is beneficial. Another important reality that must be considered by
readers is that funding and conducting multiple high-quality
randomized controlled trials (RCTs) for each of the 200 or
more individual treatment options currently available for
CLBP is simply beyond the realm of possibility. Those reviewing this special focus issue must therefore decide
whether they wish to rely on the enthusiasm of clinicians
practicing a particular treatment approach or on the skepticism of clinical epidemiologists who rely on the evidence.
A blend of both views, where possible, is perhaps the most
useful current approach.
Evidence informed versus evidence based
Ideally, there would be multiple high-quality RCTs supporting each of the interventions discussed in this special
focus issue to provide a solid EBM approach to CLBP. In
reality, SR methodology confined to high-quality RCTs
would likely find only limited evidence for many of these
interventions. Given the wealth of clinical experience
among invited authors, it was intended that articles in this
special focus issue present evidence-informed rather than
strictly evidence-based recommendations. The guiding
principle behind evidence-informed management is that authors should be aware of and use research evidence when
available, make personal recommendations based on clinical experience when it is not available, and be transparent
about the process used to reach their conclusion. The instructions to authors made it clear that articles should not
be narrative reviews founded solely on their opinions and
clinical experience. Authors were asked to systematically
search the biomedical literature to uncover, evaluate, and
summarize recent evidence using some of the methodology
recommended by the Cochrane Back Review group.  They were also given the liberty to make personal recommendations on specific aspects of a treatment in the absence of other available evidence.
All authors were asked to include a description of terminology surrounding that intervention, a detailed description
of the intervention so that patients considering a particular
intervention may know what to expect, a summary of
important historical events, the qualifications required to
administer that intervention, general information on costs
and reimbursement policies in the United States, the theories
supporting its mechanism of action, the most appropriate
indications and contraindications, the ideal CLBP patient
for that intervention, review methods used to uncover evidence of efficacy, appraisal and summary of available evidence by study design (clinical guidelines, SRs, RCTs,
observational studies [OBSs]), and discussion of known
or potential harms. Although the approach taken by each
group of authors differed somewhat, most followed this
format admirably and genuinely attempted to provide evidence-informed recommendations to assist stakeholders
evaluating these various interventions for CLBP.
Pretreatment diagnostic testing
Table 1 summarizes recommendations by the authors regarding the diagnostic testing procedures that are required
or recommended before considering each treatment approach. It should be evident to anyone reading this special
focus issue that the diagnostic testing recommended before
providing a treatment is, with few exceptions, almost identical. Every article in this special focus issue recommends
or infers that it is important to conduct a thorough history
and physical examination to rule out the possibility of serious pathology or ‘‘red flags’’ indicating organic conditions
requiring immediate attention before considering any treatment for CLBP. Certain of these articles list some of the red
flags whereas others use a more general term. None of the
articles suggest that the treatment approach should be
considered in patients with these red flags.
It is interesting to note that, despite the enormous resources devoted to daily use of diagnostic testing for CLBP,
few authors reported that such testing was required before
considering a particular intervention. Those who did suggest the use of specific diagnostic testing generally did
not support these recommendations with citations to any
studies demonstrating a change in outcomes for patients
who did and did not undergo advanced diagnostic testing
before receiving an intervention. This observation brings
into question the routine use of laboratory testing, X-rays,
computed tomography (CT), magnetic resonance imaging
(MRI), discography, nerve conduction velocity, and electromyography by clinicians evaluating CLBP.
In their review of surgical options for CLBP, Don and
Carragee discuss the failure of advanced imaging such as
CT or MRI to delineate a clear pathoanatomic cause for
a patient’s symptoms. In their discussion of epidural steroid
injections, De Palma and Slipman note that the use of
discographydwhich is often touted as superior to CT or
MRIdis controversial and has not been critically evaluated
for CLBP. This is echoed by Don and Carragee, who report
that the discography is not validated, is painful in 30% to
80% of asymptomatic subjects, and, even in a best-case
scenario, has a positive predictive value of only 50% to
60% for resolution of low back pain (LBP) after surgical
removal of the suspected pain generator identified by
A number of articles suggested that an initial trial of treatment may be used to help customize an intervention according to patient response. This was mainly noted in articles on manual therapies including spinal manipulation, mobilization, and massage, as well as trigger point injections, which also rely on manual palpation to identify areas to be injected.
Although the logic behind diagnosis by treatment appears
reasonable for these types of safe and noninvasive therapies,
there were no citations provided to support these statements.
The only two treatment approaches for which authors
cited evidence that the success of an intervention is dependent on examination findings were McKenzie methodd
which bases its treatment on findings from its customized
mechanical diagnosis and therapy assessmentdand radiofrequency neurotomy, which bases its treatments on findings from properly conducted diagnostic facet blocks.
There is clearly no consensus that commonly used diagnostic
tests hold any value in the decision-making process before
offering a treatment for CLBP.
Indications for different treatment approaches
Table 2 summarizes the proposed indications and contraindications for each of the treatment categories as reported by the authors of these review articles. One of the first observations from this table is that the indications do not differ very much and provide very little information on when
to consider a specific treatment approach. Indications for
various interventions appear to fall into one of three broad
categories: pain, nonspecific or mechanical CLBP, and
failure of other treatments.
The presence of pain is noted as the main indication for
articles on pharmacological approaches such as nonsteroidal anti-inflammatory drugs (NSAIDs), muscle relaxants, simple analgesics, opioid analgesics, adjunctive analgesics,
and various nutritional supplements. Authors frequently
noted that many of these medications are only approved
for indications other than CLBP and therefore used off label with very little or no published evidence of efficacy.
Clinicians must make a leap of faith that success noted in
conditions such as diabetic neuropathy or pain in patients
with terminal cancer can be translated to improving the
symptoms of CLBP. Although the relative advantages/disadvantages of medication classes (eg, tricyclic antidepressants vs. selective serotonin reuptake inhibitors) were discussed, very little time was spent discussing why one medication in any particular drug class should be used over its competitors. There was therefore very little guidance to clinicians or consumers as to which medication should be considered beyond the opinion and experience of the prescribing clinician. Viewed together, these articles do suggest that certain nutritional supplements, NSAIDs, simple analgesics, and muscle relaxants can be used as a first-line approach to CLBP, with the consideration of adjunctive or opioid analgesics when pain is refractory.
A number of the treatments discussed in this special
focus issue list the primary indication as nonspecific or
mechanical CLBP. These include spinal manipulation and
mobilization, massage, acupuncture, electrotherapeutic
physical modalities, prolotherapy, traction, most of the
exercise therapies, back schools, and patient education. Although vague, the description of nonspecific CLBP is
perhaps the most honest statement regarding the indication
of any intervention for CLBP given the doubts expressed
above about the usefulness of advanced diagnostic testing
for identifying the exact source of pain.
The third indication stated by authors to justify a particular procedure is that the patient has failed to respond to
other treatment approaches, which was noted in many of
the articles on injection or minimally invasive interventional procedures. This reasoning should be viewed with some
concern because it has previously been used to support dubious treatments for other health conditions which simply
cannot be cured. Although compelling, the simple fact that
previous treatments have failed is not sufficient justification
for exposing a patient to any treatment that is supported
solely by weak evidence and which is associated with considerable costs and increased risks of harms. In fact, it
could be argued that the burden of scientific proof should
perhaps be even higher for so-called rescue therapies to
prevent exposing patients to potential harms who may
simply never respond positively to any interventions.
Comparing the evidence for efficacy
Table 3 is a summary of the evidence for efficacy which
was reported by the authors in the 25 articles in this special
focus issue. This table also presents the conclusions and
recommendations based on the best available evidence considered by those authors. This table reports the number of studies discussed and does not reflect whether the studies were positive or negative for the intervention or whether the authors agreed with the results of these studies.
It is noted that the number and type of studies that were
offered to support or provide evidence of efficacy or lack
thereof varied considerably among the different treatment
categories. The only interventions that reported being included in clinical practice guidelines on LBP were back schools and brief education, NSAIDs and simple analgesics, the McKenzie method, needle acupuncture, spinal manipulation and mobilization, trigger point injections, and watchful waiting (for acute LBP).
The five interventions where the authors reported the
highest number of SRs were back schools (seven SRs), needle acupuncture (six SRs), tricyclic antidepressants (five
SRs), prolotherapy (four SRs), and traction therapy (four
SRs). The five interventions that authors reported the highest number of RCTs were needle acupuncture (19 RCTs),
spinal manipulation and mobilization (13 RCTs), lumbar
extensor strengthening exercises (11 RCTs), brief education
(11 RCTs), and epidural steroid injections (10 RCTs). The
five interventions where the authors relied primarily on
OBSsdwhich include controlled clinical trials, prospective
cohorts, and case seriesdwere Intradiscal Electrothermal
Therapy (24 OBSs), minimally invasive nuclear decompression (nucleoplasty) (10 OBSs), medicine-assisted manipulation (6 OBSs), opioid analgesics (4 OBSs),
and functional restoration (4 OBSs).
Although one might be tempted to correlate a large
number of studies with a strong level of evidence from the
scientific literature, this assumption would be an oversimplification. There were several possible reasons for reporting a high number of efficacy studies including 1) a few
studies on each of several subtypes of interventions were
combined into one broader category; 2) an intervention
has a long history of use over which more studies have been
conducted; 3) study eligibility criteria used by authors were
more lenient; 4) multiple health databases were examined
using a sensitive and comprehensive search strategy; 5)
an intervention is controversial and has attracted the interest of researchers and funders; 6) conflicting results among
studies perpetuate the need for additional research; or 7)
lack of acceptance has motivated additional research to
gain market share.
Although it appeared that studies with a higher number of
SRs and RCTs generally reported positive findings supporting efficacy, best evidence syntheses from those review
articles were often cautiously worded and offered only
lukewarm recommendations on specific comparisons (eg,
intervention vs. placebo), specific outcome measures (eg,
pain but not function), or specific follow-up periods (eg, short
term only). This may also have been a reflection of the experience and training in EBM of the authors involved, as mentioned earlier.
It was also noted that interventions discussing a high
number of OBSs seemingly did so in the absence of higher
level of evidence (eg, SRs or RCTs). Their articles also
tended to make less nuanced and more positive recommendations. These findings lend themselves to two observations
made regarding the interventions for CLBP reviewed in this
special focus issue: 1) evidence of efficacy appears less ambiguous and more positive when based mostly on OBSs;
and 2) recommendations become more restrained and conflicting when multiple SRs or RCTs are available to define
boundaries regarding the conclusions that can be drawn
from the scientific literature. In other words, the lower
the quality and quantity of available research on an intervention, the higher the enthusiasm shown by clinicians
for its efficacy. Stakeholders may wish to consider these
possibilities when evaluating different treatment approaches
Reported harms from different interventions
Table 4 is a summary of the harms (eg, minor side effects, adverse events (AEs), serious AEs, complications) reported in the 25 articles in this special focus issue, and
general estimates of their prevalence. Reported harms that
have been associated with the interventions reviewed in this
special focus issue varied considerably in nature, frequency,
and severity. Commonly reported side effects included localized pain, soreness, or discomfort, mild gastrointestinal complaints with orally ingested therapies, and vague discomforts such as fatigue, weakness, or dizziness. The reported estimated prevalence of minor and usually brief side effects varied from 1% to 76%. More serious AEs included transient or permanent disc, vertebral, neural, or spinal cord injuries, which were more commonly reported with interventions requiring injections. All were described as rare, and usually based on isolated case reports or small case series.
It is difficult to form conclusions as to the relative safety
of these interventions based on the harms reported by the
authors. Although it is tempting to assume that interventions for which numerous possible harms were reported
are inherently more dangerous that alternatives for which
no harms are listed, this does not appear to be the case.
Harmsdwhether theoretical or previously reporteddare
possible with all of the interventions reviewed in this special focus issue. The most likely explanation for this discrepancy is that those authors who put more time and
effort into searching and summarizing available evidence
regarding harms were more likely to fully report their existence. Interventions for which few or no harms were reported likely performed only a cursory search for this information, or have simply not been studied sufficiently. To fully present the risks and benefits of available alternatives during the increasingly important informed consent process, clinicians must have access to more comprehensive research and reviews of harms than those presented by most authors in this special focus issue. Additional research related to the comparative harms of common interventions for CLBP is necessary before stakeholders can even consider this aspect in their decision-making process.
This special focus issue contains review articles written
by clinicians and researchers who summarized the evidence
on 25 classes of commonly used interventions for CLBP. The wealth of information provided by these articles cannot
be understated and every article must be read in its entirety
to appreciate the particular strengths and weaknesses of the
arguments used by the authors for each treatment approach.
It is also necessary for the reader to look at the entire
special focus issue to obtain an overview of the different
treatment options and place them in perspective. Although
it was initially hoped that global recommendations regarding the use of specific interventions for CLBP could be made based on the information presented in each article, this goal has proven to elusive at this moment. When viewed as a whole, the articles in this special focus issue pose more questions than they answer. Taken together, these reviews demonstrate the serious deficiencies in the available research for many of the treatment approaches that are commonly used for CLBP because of either unavailable, insufficient, or conflicting research results. These articles do not present convincing evidence that it is currently possible to select one treatment approach over another for patients with CLBP and give very little guidance on when any specific treatment approach is indicated.
When viewed optimistically, the articles in this special focus issue do suggest that a reasonable approach to CLBP would include education strategies, exercise, simple analgesics, a brief course of manual therapy in the form of spinal manipulation, mobilization, or massage, and possibly acupuncture. In patients with longstanding or severe symptoms
and psychological comorbidities, there is some evidence
that a comprehensive multidisciplinary approach with cognitive behavioral treatment, fear-avoidance training, or functional restoration is at least as beneficial as surgery. This interpretation of the best available evidence is not materially different than the recommendations from the Practice Guidelines on Acute Low Back Pain in Adults that were published by the Agency for Health Care Policy and Research in 1994.  Although potentially heartening to the many clinicians who have adopted aspects of this approach, it is somewhat disappointing to note that 14 years after dozens of highly promoted new interventions, thousand of studies, millions of lost work days, and billions
of dollars spent on its care, so little has changed in the evidence available to guide stakeholders and support treatments for CLBP.
As noted in the review of the economic burden of LBP in
this special focus issue, the magnitude of this problem is
likely increasing in the United States and the question that
needs to be answered is whether any treatment should be
offered and widely used before there being sufficient research evidence to establish its efficacy, safety, and cost
effectiveness. It is a generally accepted principle in most
fields of health care that a treatment should not be offered
to the public until there is sufficient evidence supporting its
safety and effectiveness and a consensus by clinicians of different backgrounds as to its most appropriate indications and
contraindications. It should be evident to most readers that
this is not the norm when dealing with CLBP and additional
research is required to achieve this long-term goal. In the
interim, patients, clinicians, third-party payers, and policy
makers have a responsibility to become thoroughly familiar
with, critically appraise, compare, and openly discuss the
best available evidence presented in this special focus issue.
In this supermarket of over 200 available treatment options
for CLBP, we are still in the era of caveat emptor (buyer beware). The enthusiastic support by providers of any treatment
should be considered when reviewing available research evidence that supports its use. It is hoped that this special focus
issue will provide a starting point for stakeholders desiring
quality information to make decisions about the evidenceinformed management of CLBP.
van Tulder M, Furlan A, Bombardier C, Bouter L.
Editorial Board of the Cochrane Collaboration Back Review Group.
Updated method guidelines for systematic reviews
in the Cochrane collaboration back review group.
Bigos S, Bowyer O, Braen G, Brown K, Deyo R, Haldeman S.
Acute Low Back Problems in Adults: Clinical Practice Guideline No. 14
AHCPR Publication No. 95-0642: December 1994
Agency for Health Care Policy and Research,
Public Health Service, US Department of Health and Human Services