Orthop Clin N Am 35 (2004) 7 – 16
Diagnostic evaluation of low back pain
Eugene J. Carragee, MD, Matthew Hannibal, MD*
Stanford University School of Medicine, 300 Pasteur Drive, Room R171, Stanford, CA 94305, USA
Low back pain (LBP) is a common condition
affecting 80% of people at some point in their lives
[1]. Furthermore, LBP is the most common cause of
disability in patients younger than 45 years old, and the
medical costs, loss of work, and disability costs can
add up to at least $50 billion per year in the United
States [2]. Fortunately most LBP is benign in its course
and resolves with little or no intervention. Studies have
shown that approximately 60% of patients suffering
from LBP recover in 1 week, 90% in 6 weeks, 95% in
12 weeks, and only 1.2% remain disabled by the end of
1 year [3,4]. This natural history of nonspecific LBP
allows the physician to leave most individuals present-
ing with isolated LBP without an anatomic diagnosis
and reasonably expecting that their condition will
improve with general supportive care [5]. Certain red
flag clinical features may indicate serious underlying
conditions such as tumors, infections, or fractures. In
this subgroup an aggressive evaluation to rule out
serious underlying pathology is warranted. These red
flag features, however, are unusual. Even at the au-
thors’ university-based tertiary care practice, less than
1% of presenting patients have such an infection or
malignant pathology.
In the large majority of persons with nonspecific
LBP, nonaggressive diagnostic evaluation is recom-
mended before 6 – 8 weeks of symptoms. By then,
90% – 95% of these patients have dramatically im-
proved. It is the group remaining symptomatic at this
point that becomes a diagnostic dilemma. An evalua-
tion at this juncture may discover an occult tumor,
infection, instability, or other serious lesion. When the
* Corresponding author.
E-mail address: matthewhannibal@hotmail.com
(M. Hannibal).
0030-5898/04/$ – see front matter D 2004 Elsevier Inc. All rights reserved.
doi:10.1016/S0030-5898(03)00099-3
evaluation turns up a discitis or myeloma, the clinician
is reasonably certain the cause of the LBP illness has
been identified definitively.
Once again, however, a thorough investigation usu-
ally does not disclose such clear pathologic diagnoses.
Rather, in most cases, only common degenerative or
age-related changes usually are found. An attempt
often is made to distinguish which of these common
degenerative findings may be established firmly as
asymptomatic, minimally symptomatic, or even seri-
ously crippling in effect. Can today’s modern technol-
ogy conclusively identify the single ‘‘pain generator’’
that causes disabling LBP? To some investigators it
seems LBP illness can be so multifactorial—including
mechanical, psychologic, and neurophysiologic con-
tributors—that it is unreasonable to expect specific
anatomic study to confirm a diagnosis for every
patient’s LBP illness. Even if a pain generator is
suspected, it is not clear how this can be reliably
related to reported pain perception, impairment, and
disability in the face of complex social, emotional, and
neurophysiologic confounders (Fig. 1).
The physician evaluating severe LBP illness, when
only degenerative pathology is found, must use expe-
rience and common sense together with the current
diagnostic modalities to attain a reasonable working
diagnosis to treat the patient. It is with this situation in
mind that a discussion of the current palette of diag-
nostic entities available to the physician from history
and physical examination to discography can begin.
History and physical examination
The differential diagnosis of LBP is such that to
make an exhaustive list would be clinically impracti-
8 E.J. Carragee, M. Hannibal / Orthop Clin N Am 35 (2004) 7–16
Fig. 1. LBP Pathways.
cal. In general, however, the diagnoses can be broken
down into three major categories:
Mechanical (eg, HNP, osteoarthritis, spinal steno-
sis, spondylolisthesis, compression fracture)
Nonmechanical
Tumor: giant cell tumor, metastases, multiple mye-
loma, lymphoma
Infection: vertebral osteomyelitis, diskitis, HIV,
Lyme disease
Inflammatory arthritis: rheumatoid arthritis, DISH,
ankylosing spondylitis
Miscellaneous: osteoporosis, parathyroid disease,
hemoglobinopathies, psychosomatic disorders,
neuropathic joints
Visceral disease (ie, nephrolithiasis, prostatitis,
PID) [6,7]
Although 98% of causes of LBP may be caused by
mechanical factors, it is the other 2% caused by
malignancy, infection, visceral causes, and other ‘‘red
herrings’’ that must be considered most seriously and
are most important to diagnose quickly [1]. This begins
with a pointed history and physical examination.
The physician should note a history of smoking,
advanced age, weight loss, and a history of cancer, all
of which increase the likelihood of malignancy. Verte-
bral body infections occur most often in patients with
diabetes, a history of other infection or immunosup-
pression, drug abuse, or urogenital instrumentation. A
history of chronic NSAID use or peptic ulcer disease
may suggest the presence of a perforated ulcer with
retroperitoneal abscess causing LBP [8]. Symptoms
such as an escalating, unremitting course of pain not
improved by rest or supine positioning, night pain, and
the symptoms of cauda equina syndrome—pain, sad-
dle anesthesia, bowel or bladder incontinence—are all
important to recognize as red flags. Although severe,
incapacitating LBP that has little mechanical relief
with rest or analgesics certainly raises the concern for
spinal abscess, malignancy, or visceral disease, the
experienced clinician also recognizes these features as
characteristic of abnormal pain behavior also. When
the expedited work-up fails to turn up a catastrophic
condition, the clinician should be reluctant to assume
minor degenerative findings adequately explain the
‘‘illness.’’ The physician should assess the patient
carefully for psychologic or social factors that might
be contributing to a hyperbolic expression of LBP
perception (Fig. 2).
Once the nonmechanical causes of back pain are
ruled out, the mechanical causes can be addressed.
Back pain can often precede a herniated nucleus
pulposus for a variable length of time as the disc
material presses against and deforms the outer wall
of the annulus before its extrusion and resultant leg
 E.J. Carragee, M. Hannibal / Orthop Clin N Am 35 (2004) 7–16
Fig. 2. Psychosocial failure to accommodate normal spinal nocioception.
pain. Straining, coughing, or sneezing often worsens
this pain. Progressive back pain with onset at an
age less than 40 years presenting with morning stiff-
ness and improvement with exercise should make
one suspicious of an inflammatory or seronegative ar-
thropathy [9]. Back pain also may present in the older
individual and may be associated with exertional leg
pain representing the pseudoclaudication of spinal
stenosis. Spinal stenosis when associated with back
pain also may have a component of spondylolisthesis
that may be associated in the history with the patient
complaint of increased sharp back pains with changing
of position from sitting to standing or vice versa.
Unfortunately in clinical practice one finds that symp-
toms are usually nonspecific, rarely conform to a fixed
diagnosis, and are often overlapped or absent, making
an anatomic diagnosis that much more difficult from
history alone.
During the physical examination, particularly in
the presence of red flags by history, it is again im-
portant to rule out the serious causes of back pain by
identifying fever, abnormal blood pressure, tachycar-
dia, palpable lymph nodes, and abdominal, pelvic, or
rectal masses. When attempting to identify mechanical
causes of LBP, the examination should begin with
assessment of spine symmetry, posture, and flexibility.
The physician should note if there is increased pain
9
in extension with relief in flexion, suggesting spinal
stenosis, or increased pain with flexion, suggesting
disc disease. Palpation of the muscular structures and
spinous processes in the low back also should be
completed to help identify abnormalities and muscular
spasms. Next, a neurologic examination is performed,
including sensory, motor, and reflex testing of the
lower extremities. A straight leg test to 60° should be
performed with the understanding that it is 95%
sensitive in patients with L4 – 5 and L5 – S1 disc
herniations, but not as specific. The cross straight leg
test also should be performed, which is 25% sensitive
but 90% specific [10]. An examination of both hips,
particularly limited or painful internal rotation, should
be done to determine if arthroses are present that can
commonly mimic LBP. The Waddell signs for deter-
mination of nonorganic causes of LBP also should be
included in the standard examination [11]. Finally,
during the history and physical examination the clini-
cian should observe carefully the affect of the patient
and the context of the LBP illness as presented by the
patient. Remarkably, the objective finding of serious
emotional distress is seen more commonly in subjects
without serious underlying pathology. That is, patients
with serious spinal pathology such as chronic vertebral
osteomyelitis or unstable spondylolisthesis have been
shown to have less emotional distress despite severe
10 E.J. Carragee, M. Hannibal / Orthop Clin N Am 35 (2004) 7–16

Fig. 3. Psychological profiles in patients with chronic LBP of various causes and controls. Discogram +, patients with presumed
discogenic pain syndrome; Spondyl, patients with unstable spondylolisthesis; PVO, chronic pyogenic vertebral osteomyelitis;
soldiers/CLBP, US Army soldiers with chronic LBP and no disability; Asx Control, asymptomatic controls.

pain than subjects with nonspecific chronic LBP
illness (Fig. 3).
Once these aspects of the history and physical ex-
amination are complete, the diagnosis may be clearer
and confirmatory diagnostic testing can begin.
Imaging studies
Diagnostic imaging in LBP has a high rate of
asymptomatic abnormalities. With more sensitive
instruments this issue only becomes more problematic.
In general, therefore, imaging modalities should be
used primarily as confirmatory tests to verify or rule
out the suspected cause of LBP. There is an important
moment in the clinical encounter when the doctor
orders the test and the patient needs to be told why.
It is essential if the MRI is being ordered to rule out
malignancy or infection that the patient understands
that is the purpose. The patient should understand
clearly that the usual degenerative changes are ex-
pected but that is not what you are looking for.
Otherwise whatever findings are detailed in the radi-
ology report may be misunderstood to be the cause
of the patient’s pain, to confirm a ‘‘serious’’ condition
or prolong their disability. Often patients with nonspe-
cific low back complaints are seen for consultation,
having had an MRI and understanding that this has
discovered the ‘‘bulging anulus’’ or ‘‘dark disc’’ that
you need to treat. This misperception is almost impos-
sible to rectify after the fact and can be avoided easily
by carefully explaining beforehand what serious pa-
thology one is actually looking for.
Plain film radiography often is the first study
obtained in the imaging evaluation of the spine. Plain
radiography has the highest spatial resolution and
allows the physician to obtain information such as
the number of lumbar vertebrae, transitional variations
of the lumbosacral junction, spondylosis, deformity,
and a vacuum sign in the disc. Furthermore, with
flexion and extension radiographs, a dynamic spondy-
lolisthesis can be detected, although these are often
asymptomatic. Overall, however, plain films are con-
spicuously unrewarding in the investigation of LBP
[12]. If the reason the study is being ordered is to rule
out serious pathology or evaluate sciatica symptoms,
the MRI is a better first step.
CT scanning also has been used to evaluate the
cause of LBP but has been largely replaced by MRI.
CT can better image bony detail and is useful in those
patients who are unable to undergo MRI because of
claustrophobia or implanted metal. CT myelography,
however, remains a useful test for the determination of
 E.J. Carragee, M. Hannibal / Orthop Clin N Am 35 (2004) 7–16
bony causes of spinal stenosis. Also, a dynamic CT
myelogram may be completed that may reveal stenosis
or compression of neural elements that was not appre-
ciated on static studies. Finally, in complex deformities
the helical CT scan can be reformatted in a different
plane and may allow better anatomic understanding
and preoperative planning than MR.
MRI is the single most sensitive and most specific
investigation to reveal conditions such as disc hernia-
tion, soft-tissue or neurologic lesions, neoplasms, or
infections. In regards to LBP possibly associated with
common degenerative findings of the discs and facets,
however, it has been shown to be too nonspecific to
differentiate patients with chronic LBP from individu-
als with no LBP at all. Studies of MR findings in
asymptomatic individuals have shown 30% – 40%
have changes and these seem to increase with age
[13 – 15]. Even among symptomatic subjects, MR
findings of mild to moderate neurologic compression,
disc degeneration or bulging, and central stenosis were
not found to correlate with severity of symptoms [15].
Other work has shown only a weak association bet-
ween LBP development and even progressive degen-
erative changes seen on MR over a 5-year period [16].
High intensity zone
One MRI finding, the high intensity zone (HIZ),
has been purported to be highly specific for discogenic
LBP illness. When evaluating patients with severe
chronic debilitating back pain illness, several inves-
tigators have reported on the presence of this focus of
high T2 signal in the posterior or posterior-lateral
annulus in discs that caused pain during a subsequent
discogram. This peripheral disc high T2 signal inten-
sity, referred to as the high intensity zone (HIZ), also is
often associated with low signal intensity centrally in
the disc. Aprill and Bogduk reported a prevalence of
HIZ in the mid posterior annulus of 28% of patients
with severe LBP considering fusion. On discography
they found that the HIZ correlated well with the
presence of complete annular penetration of the dye
and concordant pain reproduction with injection. Of
40 HIZ-positive discs, 38 had ‘‘positive’’ discography,
whereas only 22 of 78 discs with positive injections
were HIZ negative. In analyzing their data, Aprill and
Bogduk calculated the positive predictive value of HIZ
in diagnosing a symptomatic disc to be approximately
90% [17]. Other investigators, however, have reported
on the presence of these high-signal zones of the disc
annulus including in asymptomatic subjects. Jensen
et al noted a 14% incidence of bright annular fissures in
11
100 asymptomatic subjects undergoing lumbar MRI.
Buirski and Silberstein reported high-intensity annular
fissures in 31% of control subjects and 47% of symp-
tomatic subjects. These investigators questioned
whether the HIZ finding could be used to predict
symptomatic ‘‘discogenic’’ LBP. Schellhas et al ex-
amined the lumbar MRI of 17 asymptomatic subjects
and found HIZ in only 1 patient (6%).
The authors’ group at Stanford University exam-
ined the prevalence of HIZ in LBP patients and
asymptomatic volunteers undergoing discography in
an experimental setting. The prevalence of an HIZ was
59% in the symptomatic group and 24% in the asymp-
tomatic group. In the symptomatic group, 72.7% of the
discs with an HIZ were positive on discography,
whereas only 38.2% of the discs without an HIZ were
positive. In the asymptomatic group, however, 69.2%
of the discs with an HIZ were also painful on discog-
raphy, whereas 10% of the discs without an HIZ were
painful. That is, the rate of positive pain response with
disc injection was the same in HIZ discs of sympto-
matic patients and asymptomatic volunteers.
In summary, although an HIZ may predict the
presence of an annular fissure, the finding is not
pathognomonic for discogenic LBP illness, because
many asymptomatic subjects have the finding.
Diagnostic anesthetic injections
Diagnostic anesthetic blockade may be used for
diagnosis or for therapy. Neural blockade as a diag-
nostic tool in painful conditions of the spine has be-
come a popular approach to distinguish clinically
relevant from irrelevant changes seen on imaging
studies. The use of these injections for diagnostic
purposes rests on three premises. First, the pain gen-
erator is a single discrete entity located in an exact
location and the pain travels by way of a unique and
consistent neural pathway without significant central
modulation. Second, the anesthetic blockade affects
only the intended pain generator and has no overlap
with other nerves or tissues. Third, the relief experi-
enced is attributable solely to the specific site block of
the afferent neural pathway and is unaffected by cen-
tral processing (placebo effect) or motivational report-
ing. The validity of these assumptions is limited by the
complexities of pain perception and the effect of local
anesthetics. It is unlikely that even a small minority of
patients meet these criteria for validity [18]. Studies
have shown that even without a placebo effect, one
does not have to block the actual painful site of pa-
thology directly to have subjective pain relief [19,20].
12 E.J. Carragee, M. Hannibal / Orthop Clin N Am 35 (2004) 7–16
Pain relief had been reported with blocks distal to a
lesion (such as a herniation disc), adjacent to regional
pathology, and even with central stimulation.
Despite these known neurophysiologic limitations,
these injections commonly are used for the diagnosis
of suspected pathology in the facet joints or the
sacroiliac (SI) joints.
Facet injections can be intra-articular injections or
medial branch blocks. Both of these injections assume
the LBP is caused directly by irritation in the facet
joints. Some investigators, with limited evidence and
no gold standard, suggest the prevalence of facet joint
involvement in LBP seems to be 15% – 40%, with back
pain caused solely by the facets as low as 7% [21 – 23].
Standard blockade injections of the medial branches
seem to anesthetize the joint and also the muscles,
ligaments, and periosteum that they innervate. Other
investigators suggest the reproducibility of the single
(uncontrolled) injection is not high, and the specificity
may be approximated at only 65% [24]. Furthermore,
in one study 30% of individuals receiving subcutane-
ous saline injections rather than lumbar facet joint
blocks experienced relief of their facet joint pain [21].
In another study it was concluded that pain relief after
facet joint blocks do not correlate with radiographic
evidence of facet arthrosis [25]. Some nerve ablation
procedures have shown more promise, but there
remains no standard test with which to establish the
validity of facet blocks of any type in making a
diagnosis [26]. No method thus exists to confirm that
facet joint pain per se is diagnosed regularly and
accurately by injection blockade as a primary cause
of LBP in most patients.
The sacroiliac (SI) joint also can be the source of
LBP. Injection of the sacroiliac joint in subjects with-
out pain can produce pain in the buttock, posterior
thigh, and knee [27]. Other studies have been unable to
find a real correlation between the physical examina-
tion diagnosis of SI joint pain and relief after an SI joint
injection [28]. It is still not clear to what extent the
anesthetic blockade needs to actually infiltrate the
joint, because injections into the SI ligaments also
have been effective. Their effectiveness, however, has
not been confirmed in controlled studies. In general,
patients who respond to SI joint injections with pain
relief do not have a characteristic history or response to
physical examination maneuvers. Also, no imaging
study finding can determine predictably a response to
an SI joint anesthetic injection. All these factors make
the definitive identification of SI joint pain difficult; it
is best done with correlation of physical examination,
imaging studies, and response to injection. Except in
the presence of clear pathology (tumor, fracture, in-
fection), the diagnosis of SI joint pain in patients with
nonspecific histories is at best tentative with limited
scientific basis.
Discography
Discography is used commonly in the diagnosis of
nonspecific LBP when common degenerative changes
are suspected of causing clinically disabling disco-
genic pain. With the advent of MRI and the increased
sensitivity of the test to degenerative changes, the
demonstration of pathology at multilevel disc has
increased demand for a provocative spinal test, such
as discography, that purports to discriminate between
truly symptomatic changes seen on MRI and simply
benign disc changes. Furthermore, even with the high-
est resolution MRI scan, some lumbar disc pathology
(such as occult annular disruption) may escape detec-
tion completely. Discography, therefore, sometimes is
done in the lumbar region with or without obvious disc
degeneration on preliminary imaging studies. In one
study with simple discography techniques, disc
injections seemed to improve modestly both surgical
outcomes [29]. Madan et al found that after circum-
ferential fusion, clinical outcome rates with and with-
out discography were not statistically different. It
follows that provocative discography is poorly under-
stood and has limited efficacy in improving clinical
outcome scores after low back surgery for discogenic
back pain [30].
The suggested clinical indications for discography
are wide ranging and highly individualized. The most
common use of discography, however, is to determine
whether degeneration within a disc seen on imaging
studies is the primary clinically significant source of a
patient’s LBP illness. Prospective, controlled inves-
tigations of discography have demonstrated some key
points about the study. First, anatomically normal discs
at discography are not painful when injected at low
pressures with saline or dye. Second, discs that prove
to be intensely painful and clinically concordant when
injected are more likely to have annular fissures into or
through the outer third of the disc annulus.
Discography in asymptomatic subjects
Discography has been controversial since its in-
ception in 1948. Holt and later Massie and Stevens
questioned whether disc injections are not often pain-
ful even in subjects without clinical LBP illness. Walsh
et al [31] found only 10% of young, healthy men
without an LBP history and with little degenerative
disk disease (DDD) had significant pain intensity with
 E.J. Carragee, M. Hannibal / Orthop Clin N Am 35 (2004) 7–16
injection. More recently the use of discography in cer-
tain subgroups of patients, especially emotionally dis-
tressed or pain sensitized individuals, has been called
into question. Multiple studies by Carragee et al have
suggested the test may be an unreliable indicator of a
chronic LBP (CLBP) patient’s primary cause of illness
[32 – 37]. The first problem arises in attempting to
define what the criteria are for a positive test. The next
issue is the determination of the specificity of the
positive test, that is, how frequently a positive test
correctly identifies discogenic pain as the primary
cause of CLBP illness. In the absence of a gold stan-
dard to confirm the diagnosis of primary discogenic
LBP illness, the question of definition criteria and
specificity must be approached indirectly. To address
these questions a series of empiric and experimental
studies were done to evaluate critically the validity
of discography.
The first step was to define a positive disc on
discography. There is currently no consensus in the
spine community of what constitutes a positive disc
injection. At a minimum, according to Walsh et al,
a positive test in a symptomatic patient would include
a negative control disc, concordant or usual pain pro-
duction during injection, demonstration of pain be-
havior with injection, and at least a 3 out of 5 or a 6 out
of 10 pain intensity. For an experimental asympto-
matic subject, Walsh et al defined a false positive test
as all of the above but without a ‘‘concordancy’’ as-
sessment. The first experimental series using this
protocol was done by Walsh et al in 1990, as described
previously, and demonstrated 10% of that selected
group (healthy, young men without much DDD) has
at least one disc with pain production at a 3 out of
5 level (‘‘bad’’ pain). The authors’ group at Stanford
University then performed another trial using the same
protocol on 26 subjects with no prior history of LBP
[32]. These patients were broken down into three
groups: pain free, chronic pain of nonlumbar origin,
and somatization disorder. The results showed that the
intensity of the pain reported by the subjects was
predicted by the presence of annular disruption, any
chronic pain syndrome, and abnormal psychometric
testing. It was found that 20% of these subjects had at
least one positive injection despite having no history
of LBP. In this group, seven patients were involved in
disability and compensation claims and six (86%) of
these had at least one positive disc. This suggests that
false-positive results may be more likely in the pa-
tients with chronic pain processes, emotional distress,
and litigation.
The next study, done by the Stanford University
group, looked at the rate of painful disc injections in
previously operated lumbar discs: 47 patients, all
13
status post a lumbar microdiscectomy [33]. These
patients were broken into symptomatic and asymp-
tomatic groups (for LBP) and all patients were sub-
jected to MRI and provocative lumbar discography.
The study found that up to 40% of asymptomatic
patients with normal psychometric testing had a pain-
ful disc on injection and that this is not statistically
different from psychologically normal symptomatic
patients after the lumbar microdiscectomy. Abnormal
psychometric testing and chronic pain, however, was
associated with an increased rate of painful injections
in symptomatic patients.
The specificity of concordancy rating in
discography
From these studies it seems clear that some sub-
jects with disruption of the outer annulus, whether
symptomatic or not at the time of injection, experience
pain during discography. Furthermore, those with
chronic pain or emotional distress are far more likely
to report significantly painful injections than those
without these risk factors, which is probably not
surprising. The reliability of the positive test would
then seem to critically hinge on the concordancy re-
port to exclude painful injections that are not true
positive tests. In the experimental design a positive
injection could not include concordancy in an asymp-
tomatic subject.
To address the issue of concordancy one must
determine if a patient is actually capable of differen-
tiating pain elicited from a disc during discography
from pain caused by other sources. Three studies were
done to help elucidate this issue. The first, in 1997,
involved a case-report series of patients who were
diagnosed with discogenic LBP on the basis of disco-
gram but who were shown subsequently to have non-
spinal causes for their pain [35]. The pain generators
were found to be pathology in the SI joint in two
patients and a neoplasm in one. The patient’s usual
symptoms were relieved once the offending nondisc
pathology was addressed. In each case the patient had
described the pain with injection as reproducing the
usual pain originating from a nondiscogenic source.
This study established that the injection of a normally
painless disc can reproduce the sensation from dis-
tant pathology.
A follow-up on an experimental study was done
involving eight subjects with no history of low back
symptoms and normal psychometric results who were
to undergo posterior iliac crest bone graft harvesting
for reasons other than lumbar surgery [36]. Two to four
months after the surgery these subjects had multilevel
14 E.J. Carragee, M. Hannibal / Orthop Clin N Am 35 (2004) 7–16
discography. The results showed that four of the eight
patients (50%) had a single positive discogram result
with severe pain and symptoms concordant with the
iliac bone graft site or the surgical site. Again, the
presence of annular tears predicted the presence of
concordant pain production.
In the last study discography was performed in
25 volunteer subjects with no clinical back pain illness
but who had persistent LBP not associated with physi-
cal restrictions [37]. In 36% of these subjects with
common backache, one or more discs were signifi-
cantly painful and concordant. This study showed that
clinically insignificant pain, possibly from a disco-
genic source, can have a fully painful and concordant
result with discography that may be confused with
clinically significant pathology elsewhere.
The overall result of these studies suggests that
balancing the discography result on the basis of a
subjective pain quality assessment may be problematic
in many patients and certainly has not been proven to
be reliable in a clinical or experimental setting. Judg-
ing from these data, it is not clear what concordancy
means when it is elicited. Sensory input from the
lumbar and pelvic regions is not likely to have suffi-
ciently detailed sensory cortical representation for a
patient to reliably distinguish discogenic from non-
Box 1. Practical use guide for discography
 Best case
1. Negative discogram (next to other
pathology—eg, spondylolisthesis)
2. Positive single level, normal
psychologic status, normal social
profile (no Worker’s Compensa-
tion, litigation, secondary gain)
 Unclear or doubtful usefulness
1. Two-level positive, normal
psychologic status, normal
social profile
2. Postoperative discs, normal
psychologic status, normal social
profile
3. Intermediate (at risk)
psychometrics, single-level
 Poor usefulness
1. Spine with multilevel pathology
2. Abnormal or chronic behavior
3. Abnormal psychometric findings
4. Disputed compensation cases
discogenic pain or from one level to another. The
failure to confirm a highly reliable concordancy rate
in these tests has led the authors to re-evaluate the use
of discography as a highly reliable means of diagnos-
ing discogenic LBP.
Given these considerations, a practical guide to
clinical discography can be made (Box 1).
Summary
The diagnostic evaluation of chronic LBP is at best
a complex and involved undertaking. The most im-
portant part of the process lies in the knowledge of the
patient and a solid history and physical examination.
From there, most of the serious and life-threatening
causes of LBP can be elucidated and studies may be
used for confirmation. Imaging studies are used most
practically as confirmation studies once a working
diagnosis is determined. MRI, although excellent at
defining tumor, infection, and nerve compression, can
be too sensitive with regard to degenerative disease
findings and commonly displays pathology that is not
responsible for the patient’s symptoms. As an exam-
ple, the high-intensity zones (HIZ) seen on MRI are
reliable in determining annular defects in the disc but
are not reliable in establishing internal disc disruption
as the cause of LBP.
Discography is the primary tool used by many
physicians to determine the true pain generator when
discogenic LBP is suspected. Because the reliability of
the patient response is fundamental to discography, in-
terpreting the test in different settings must be consid-
ered. In individuals with disc degeneration and annular
defects, discography may elicit LBP with injection
whether the patient is symptomatic with serious LBP
or not. The pain response may be amplified in those
subjects with issues of chronic pain, social stressors,
such as secondary gain or litigation claims, or psycho-
logic distress disorder. These factors have been shown
experimentally to be associated with an increased risk
for a false positive injection. The ability of an individ-
ual to differentiate the true site of LBP by the quality of
sensation with disc injection (concordancy) of pain
produced by the injected disc also may not be reliable.
In fact, individuals may not have the neural discrim-
ination to differentiate sclerotomal pain originating
from different sites in the low back and pelvis.
One may realize that chronic LBP illness may not
stem from a mechanical spinal disorder alone. In fact,
the mechanical pathology may be just a portion of the
problem with amplification by neurophysiologic, so-
cial, and psychologic issues. Chronic disabling LBP
commonly is confounded by chronic pain, emotional
 E.J. Carragee, M. Hannibal / Orthop Clin N Am 35 (2004) 7–16
troubles, poor job satisfaction, alcohol and narcotic
abuse, and compensation issues, just to identify a few.
It would follow that expecting to identify a single
cause for this symptom complex is impractical and
any single test may not be a reasonable approach.
Furthermore, surgical correction of the mechanical
portion of chronic LBP, even if correctly identified,
then can be expected only to relieve a portion of a pa-
tient’s symptoms as long as the confounding issues
continue to be significant or have become lifelong
adaptive mechanisms. In the end, the discogram and
other diagnostic tests are tools that have clear limita-
tions. In this field, clinical judgment begins and ends
with an understanding of a patient’s life and circum-
stances as much as with their specific spinal pathology.
References
[1] Wheeler AH. Diagnosis and management of low back
pain and sciatica. Am Fam Phys 1995;52:1333 – 41.
[2] Wipf JE, Deyo RA. Low back pain. Med Clin N Am
1995;79:231 – 46.
[3] Lawrence RC, Helmick CG, Arnett FC, et al. Estimates
of the prevalence of arthritis and selected musculo-
skeletal disorders in the United States. Arthritis Rheum
1998;41:778 – 99.
[4] Epidemiology of back pain. In: Borenstein DG, Wiesel
SG, Boden SD, editors. Low back pain: medical diag-
nosis and comprehensive management. Philadelphia:
WB Saunders; 1995. p. 22 – 7.
[5] White AAD, Gordon SL. Synopsis: workshop on idio-
pathic low back pain. Spine 1982;7:141 – 9.
[6] Hicks GS, Duddleston DN, Russell LD, et al. Low back
pain. Am J Med Sci 2002;324(4):207 – 11.
[7] Hall S, Bartesson JD, Onofrio BM, et al. Lumbar spinal
stenosis. Ann Intern Med 1985;103:271 – 5.
[8] Kage B. Medical aspects of low back pain. Sem Spine
Surg 2000;12(2):60 – 6.
[9] Deyo RA, Rainville J, Kent DC. What can the history
and physical examination tell us about low back pain?
JAMA 1992;268:760 – 5.
[10] Deyo RA, Loeser JD. Herniated lumbar intervertebral
disk. Ann Intern Med 1990;112:598 – 603.
[11] Waddell G, McCulloch JA, Kummel E, et al. Non-
organic physical signs in low back pain. Spine 1980;
5:117 – 25.
[12] Jarvik JJM, Hollinworth W, Heagerty P, et al. The lon-
gitudinal assessment of imaging and disability of the
back (LAIDBack) study—baseline data. Spine 2001;
26:1158 – 66.
[13] Jensen M, Brant-Zawadzki M, Obuchowski N, et al.
Magnetic resonance imaging or the lumbar spine in
people without back pain. N Engl J Med 1994;331:
69 – 73.
[14] Boden S, Davis D, Dina T, et al. Abnormal magnetic
resonance scans of the lumbar spine in asymptomatic
15
subjects: a prospective investigation. J Bone Joint Surg
[Am] 1990;72A:403 – 8.
[15] Beattie P, Meyers S, Stratford P, et al. Associations be-
tween patient report of symptoms and anatomic impair-
ment visible on lumbar MR imaging. Spine 2000;25:
819 – 28.
[16] Elfering AP, Semmer N, Birkhofer D, et al. Risk factors
for lumbar disc degeneration—a 5-year prospective
MRI study in asymptomatic individuals. Spine 2002;
27:125 – 34.
[17] Aprill C, Bogduk N. High-intensity zone: a diagnostic
sign of painful lumbar disc on magnetic resonance
imaging. Br J Radiol 1992;65(773):361 – 9.
[18] Hogan QH, Abram SE. Neural blockade for diagnosis
and prognosis. Anesthesiology 1997;86:216 – 41.
[19] Siddall P, Cousins M. Spinal pain mechanisms. Spine
1997;22:98 – 104.
[20] Kibler R, Nathan P. Relief of pain and paresthesias
by nerve block distal to a lesion. J Neuro Neurosurg
Psychiatry 1960;23:91 – 8.
[21] Schwarzer AC, Wang S, Bogduk N, et al. Prevalence
and clinical features of lumbar zygapophyseal joint
pain: a study in an Australian population with chronic
low back pain. Am Rheum Dis 1995;54:100 – 6.
[22] Jackson RP, Jacob RR, Montesano PX. Facet joint in-
jection in low back pain: a prospective statistical study.
Spine 1988;13:966 – 71.
[23] Carrette S, Marcoux S, Trouchon R, et al. A controlled
trial of corticosteroid injection into the facet joints
for chronic low back pain. N Engl J Med 1991;325:
1002 – 7.
[24] Schwarzer AC, Aprill CN, Derby R, et al. The false
positive rate of uncontrolled diagnostic blocks of the
lumbar zygapophyseal joints. Pain 1994;58:195 – 200.
[25] Schwarzer AC, Wang SC, O’Driscoll D, et al. The abil-
ity of computerized tomography to identify a painful
zygapophyseal joint in patients with chronic low back
pain. Spine 1995;20:907 – 12.
[26] Destouet JM, Gilula LA, Murphey WA, et al. Lumbar
facet joint injection: indications, technique, clinical cor-
relation, and preliminary result. Radiology 1988;145:
321 – 5.
[27] Hildebrant J. Relevance of invasive techniques in treat-
ing and diagnosing chronic low back pain: an anes-
thesiologic perspective. Sem Spine Surg 2000;12(2):
67 – 73.
[28] Dreyfuss P, Michaelson M, Pauza K, et al. The value
of medical history and physical examination in diagnos-
ing sacroiliac pain. Spine 1996;2594 – 602.
[29] Colhoun E, McCall IW, Williams L, et al. Provocation
discography as a guide to planning operations on the
spine. J Bone Joint Surg [Br] 1988;70:267 – 71.
[30] Madan S, Gundanna M, Harley JM, Boeree NR, Samp-
son M. Does provocative discography screening of dis-
cogenic back pain improve surgical outcome? J Spinal
Disord Tech 2002;15(3):245 – 51.
[31] Walsh TR, Weinstein JN, Spratt KF, et al. Lumbar dis-
cography in normal subjects: a controlled, prospective
study. J Bone Joint Surg [Am] 1990;72:1081 – 8.
16 E.J. Carragee, M. Hannibal / Orthop Clin N Am 35 (2004) 7–16
[32] Carragee EJ, Tanner CM, Khurana S, et al. The rates of
false-positive lumbar discography in select patients
without low back symptoms. Spine 2000;25:1373 – 80.
[33] Carragee EJ, Paragioudakis SJ, Khurana S. 2000 Volvo
award winner in clinical studies: lumbar high-intensity
zone and discography in subjects without low back
problems. Spine 2000;25(23):2987 – 92.
[34] Carragee EJ, Chen Y, Tanner CM, et al. Provocative
discography in patients after limited lumbar discec-
tomy. Spine 2000;25(23):3065 – 71.
[35] Carragee EJ, Tanner C, Vittum D, et al. Positive provo-
cative discography as a misleading finding in the evalu-
ation of low back pain. NASS Proceedings. New
Orleans, 1997. p. 388.
[36] Carragee EJ, Tanner CM, Yang B, et al. False-positive
findings on lumbar discography. Spine 1999;24(23):
2542 – 7.
[37] Carragee EJ, Alamin TF, Miller J, et al. 2001 out-
standing paper award. Provocative discography in vol-
unteer subjects with mild persistent low back pain.
Spine J 2002;2:25 – 34.