Lumbar pain

1. Precision Diagnosis of Spinal Pain
Nikolai Bogduk, MD, PhD,FAFRM
Faculty of Medicine, University of Newcastle, Newcastle, Australia
Cervical Spine Research Unit, Mater Misericordiae Hospital, Newcastle,Australia
Richard Derby;MD
Spine Care, San Francisco, California, USA
. ' Charles Aprill, MD
Magnolia Diagnostics, New Orleans, Louisiana, USA
Susan Lord, BMed, PhD
Faculty of Medicine, University of Newcastle, Newcastle, Australia
Cervical Spine Research Unit, Mater Misericordiae Hospital, Newcastle, Australia
Anthony Schwarzer MB BS, PhD, FRACP
Faculty of Medicine, University of Newcastle, Newcastle, Australia
EDUCATIONAL OBJECTIVES *
The general objective of the course .is to replace .
pattern recognition and habit with an evidence-based,
problem-solving approach for the investigation and man-
agement of spinal pain. Specifically participants will gain
an understanding of
1. . the anatomicalbasis of spinal pain
2. the spectrum of lesions or disorders that commonly
underlie spinal pain
3. the procedures available for pin-pointing sources of
spinal pain
4. the validity ofthese diagnosticprocedures
5. the epidemiology of various commondisorders
6. the therapeutic utility ofprecision diagnosis
7. the merits of a structural organic approach to spinal
pain comparedto the classical biopsychosocial model .
of spinal pain.
INTRODUCTION
Spinal pain may be localized or it may be associated.
with somatic referred palm It can be described in terms of
the anatomical regions in which it is perceived (Merskey
and Bogduk 1994), The most common^forms of spinal pain
are lumbar spinal pain (back pain), cervical spinal pain
(neck pain), headaches which are due to referred pain from
the neck, and sacral spinal pain. Although there are other
regional spinal pain problems, it is these that have been
most fully and reliably explored. Similarly, although pro-
cedures have been developed for the investigation of tho-
racic spinal pain (Chua and Bogduk 1995; Dreyfuss et al. -
1994a; Schellhas et al, 1994) -and the upper cervical
synovial joints (Dreyfuss et al. 1994b) their validity and -
utility have not been determined; nor has the epidemiology
of the conditionsthey purportedly diagnose.
It is quite clear from clinical experience and from
formal studies that when a patient presents with spinal
pain, there are no clinical features that permit the sourceof
pain to be diagnosed (Barnsley et al. 1995; Schwarzer et'
al. 1994a,b; 1995aIb)c). Even imaging studies do not pro-
vide a diagnosis. So-called "red flag" conditions are indi-
cated by elements of history, including the patient's age,
past history, and general medical health. Tumors, infec-
tions, systemic diseases, inflammatory disorders and vis-
ceral diseases can and should be overtly suspected on the
basis of history, clinical examination and serology, and do
not requireimaging as screening tests (Bogduk 1995;Deyo
and Diehl 1986).
Once medical conditions have been excluded, imag-
ing studies and EMG are immaterial for determining a
diagnosis. The habit has been to rely on EMG and CT or
MRI to detect disc herniation, but this condition presents
with leg pain and objective neurological signs. It is illogi-
cal and wasteful to order investigations designed to detect
disc herniation when there are no grounds clinically for
suspecting this condition; and more patently so when disc
herniation is obviously not the cause of pain.
For a patient with spinal pain without neurological
signs, the question, is: From where does the pain stem;
what is its source? EMG, CT, and MRI do not answer this
313

2. question. The appropriate investigations are ones that do
answer this question.
ANATOMICAL BASIS
For a structure to be a source of pain it must have a
nerve supply. In the neck and in the lumbar spine, the dor-
sal rami of the spinal nerves supply the 2ygapophysial
joints and the posterior spinal muscles (Bogduk 1982;
Bogduk et al. 1982). The intervertebral discs are supplied
by anterior and posterior longitudinal plexuses fed by the
sympathetic trunks and sinuvertebral nerves (Bogduk et al.
1981;-Bogduket~al. L9SS; Green etal. 1990; Mendel et al. .
• 1992). The innervation of the sacroiliac joints is still in ..
dispute; it is not clear whether these joints are supplied
from the front by branches of the rumbosacral ventral rarni
and from the backby-sacral dorsal rami or solely from the_
back (Grot et al. 1995). " '
The zygapophysial joints, the intervertebral discs and
the sacroiliac joints are the cardinal candidates for sources.
of chronic spinal pain, and it is these structures that"have "
attracted the greatest deal of scientific attention'in'rec'ent "
years. Although suspected and proclaimed as sources of
pain, the spinal muscles and spinal dura have not attracted
the same degree of scientific scrutiny. They remain possi-
ble sources of pain (Bogduk 1992)but for lack of evidence ..
they are not considered further in this course.- • - - . . -
SOURCES OF SPINAL PAIN
It is quite clear from experiments in normal volun-
teers that spinal pain and referred pain can be produced by
stimulating the lurribajr zygapophysial joints (McCall et al.
1979; Mooney and Robertson 1976), the cervical zygapo-.
physial joints (Aprill et al. 1990; Dwyer et al. 1990) and
the sacroiliac joints (Fortin et al. 1994). Normal interver-
tebral discs do not hurt in asymptomatic volunteers but
they are clearly sensitive to provocation in certain patients
with spinal pain (Bogduk 1994a; Bogduk et al. 1995a).
Although the zygapophysial joints can be a sourceof
spinal pain, there is no singular nor common lesion that
renders them painful and which can be detected by imag-
ing. Lesions such as small fractures, meniscal tears and
hemarthroses have been detected in post-mortem studies,
and have been shown to be invisible in plain radiographs
(Jonsson et al. 1991; Taylor et al. 1990; Twomey et al.
1989); but it has still to be determined if such lesions un-
derlie backpain and neck pain stemming from thezygapo-
physial joints.
Neither in the neck nor in the back does osteoarthro-
sis seen on X-ray correlate with the presence" of pain
(Friedenberg and Miller 1963; Magora and Schwartz
1976; Lawrence et al. 1966); nor do CI features correlate
with proven lumbar zygapophysialjoint pain (Schwarzer et
al. 1995d). Consequently, in the pursuit of zygapophysial
joint pain all that can be done to test if a suspected joint is
the source pain is to anaesthetize that joint. Similarly,
there are no demonstrable lesions that identify a painful
sacroiliac joint. Like the zygapophysial joints that joint
must be anaesthetized to show if it is painful or not.
Intervertebral discs are difficult to anaesthetize. Intra-
discal injections of local anesthetic may succeed in reliev-'
irig a patient's pain, but such injections are liable to false-
negative results if the injected agent fails adequately to
infiltrate the nerve endings in the outer anulus fibrosus
that mediate the patient's pain Consequently, the best
available criterion standard for discogenic pain is disc
stimulation, in which a needle is used to distend the disc in
the course_pf_discography, in an effort to reproduce the
patient's.pain. ._ __ JZ~"1~"'
Although no radiologically detectable lesions "have""
been found tojffect the synovial joints of the spine, this is .
.not the case for painful lumbar discs. The studies of Van- -
haranta et al. (1987), subsequently reappraised "byMoneta"
et al. (1994), show a clear and statistically significant cor-
relation between disc pain and grade 3 fissures of the anu-
lus' fibrosus (Table 1) (Fig. 1). Moreover,.-•.logistic":;:
* regression has shown that anulax fissures alone correlate~
with disc pain; disc degeneration expressly does not
(Moneta etal. 1994).
Table 1. The correlation between anual disruption and
reproduction ofpain bydisc stimulation. "
Anular Disruption.
Pain Repro-
duction
Exact
Similar
Dissimilar
None
GDE3
43
32
9
16
GDE2
29
36
• 11
24
GDE1
6
21
6
67
• ...
GDEO
4
8
2
86
Based on Moneta et al. 1994.
X2=H8;p<0.001.
Anular fissures are the characteristic feature of inter-
nal disc disruption (TDD), an entity rendered evident by
CT discography, and most likely traumatic in origin
(Bogduk 1991; Bogduk et al. 1995a). In some cases, pain-
ful fissures are evident on MRI (Aprill and Bogduk 1992)
(Fig. 2), but disc provocation is nonetheless required to
prove conclusively that the apparently affected disc is, in-
deed, painful.
PROCEDURES
The zygapophysialjoints of the neck and back can be
anaesthetized under image-intensifier guidance by injec-
tions of local anesthetic either into the target joint or onto
the medial branches of the dorsal rami that supply them
(Bogduk et al. 1995b,c) (Figs. 3 and 4). The sacroiliac
joint can be anesthetized by intra-articular injection of lo-
cal anesthetic (Bogduk et al. 1995b; Schwarzer et al.
1995b) (Fig. 5).

3. c .TUT
ILL
ISCO
NEW ORL
26-AUG-
18 = 36
ASN 1538
SEQ NO
TBL 97
SPINE
12.7
T 16.1
Fig. 1. CT scan of an L4-5 disc with a grade 3 anular fissure.
KU 1
SEC
?CENT 283
?HIND 185
Cervical and lumbar discs can be stimulated by nee-"-
dies inserted into the nucleus of these discs (Fig. 6). Con-
trast medium or normal saline can be injected carefully to
distend the disc in an effort to reproduce the patient's pain
(Bogduketal. 1995a).
All of these procedures require skill, care and good
facilities in order to avoid possible complications (Bogduk
et al. 1995a,b). Patients,must be carefully briefed and in-
formed in order to obtain reliable information;procedures
performed roughly or without precision compromise the
quality of information obtained and its reliability.
VALIDITY
SYNOVTAL JOINT BLOCKS
Placebo responses are the greatest threat to the valid-
ity of synovial joint blocks. Formal studies have shownthat
some 30% of patients ostensibly undergoing lumbar zyga-
pophysiai joint blocks can report complete relief of their
back pain following simply a subcutaneous injection of
normal saline (Schwarzer et al. 1995a). Furthermore, sin-
gle diagnostic blocks are unreliable. They are associated
with a false-positive rate of 32% in the lumbar spine
(Schwarzer et al. 1994c) and 28% in the cervical spine
(Barnsley et al. 1993a); patients who respond to an initial
block will not necessarily respond to a confirmatory block.
For diagnostic blocks to be valid they must be con-
trolled in each and even' patient. Without controls, two out
of every three apparently positive lumbar zygapophysial
joints will be false-positive (Schwarzer et al. 1994c). Fail-
ure to implement such controls invalidates all of the previ--
ous literature on these blocks (Bogduk et al. 1995b,c).
Controlled diagnostic blacks must become the standard of
care lest the use of these blocks continue to be invalid.
Two forms of control are available: saline and
comparative local anesthetic blocks. Saline controls under
double-blind conditions are rigorous but impractical. They
require at least three procedures. The first must be per-
formed using a local anesthetic in order to identify a puta-
tively painful joint. (There is no sense in blocking'a joint
saline when it might not even be painful.) The second
block cannot routinely be normal saline for would-be mis-
chievous patients would know that the second block would
315

4. Fig. 2. An MRI scan of the disc in Pig. 1, showinga high in-
tensity zone lesion in the posterior anulus of the L4-5 disc.
ahvays be the dummy. To maintain, chance, a second and a
third block must be implemented each randomly deter-
mined to be either a local anesthetic or saline. Becauseof
the logistics involved in performing,triple.blocks,their _use
is likely to remain restricted to researchunits or fastidious
spine centers.
Comparative local anesthetic blocks conveniently cir-
cumvent the logistic problems and ethical problems of us-
ing saline controls. Their underlying principle is that
genuine patients can be identified if on separate occasions
they obtain long-lasting relief when a long-actingagent is
used and short-lasting relief when a short-acting agent is
used. The construct validity of this paradigm has been
formally established (Barnsley et al. 1993b). Moreover,the
paradigm has been tested against placebo and found to be
robust (Lord et al. 1995a).
However, comparativeblocks are not totally reliable.
On the one hand, some 15% of patients who satisfy the
criteria for a positive response also respond to placebo
Fig. 3. A lateral radiograph of the cervical spine showingnee-
dles in position for the execution of C4,5 medial branch blocks. ,-
(Lord et al. 1995a). Thus, comparative blocks are only
85% reliable. On the other hand, some patients exhibit
enigmatic responses to comparative blocks, e.g. prolonged
response to lignocaine, and would fail to be called positive
responders. However, 65% of such patients- withstand-
challenge with placebo (Lord et al. 1995a). Their enig-
matic responses are not placebo responses, and probably
reflect the affect of local anesthetics on dynamic, as op-
posed .to closed, sodium channels (Butterworth and
Strichartz 1990). •
Clinicians, therefore, need to be fluid with respect to
the criteria for positive responses to comparative blocks.
Calling positive all patients who respond to both blocks
irrespective of duration of response increases the sensitiv-
ity but lowers the specificity of the blocks. Retaining strict
criteria of differentiating long-acting and short-acting
agents, increases the specificity but at the cost of sensitivity
(Lord et al. 1995a).

5. Fig. 4. /An oblique radiograph of the lumbar spine showing a needle in position for the execution of anL3 medial branch,block. Contrast ~ '
medium shows how focal the deposition, is of the injectate. - - - •—... - ' -"=-.-^-;-_..
t. f
Fig. 5. Anantero-posterior radiograph of a sacroiliac arthrogram.
317

6. Fig. 6. Anterior and lateral vievs of an L5-S1 disc before and after injection of contrast medium in the conduct of disc stimulation and
discography.
applied, liberal criteria for comparative blocks can be tol-
erated; but if neurodestructive procedures are to be under-
taken, stricter criteria or even saline controls should be
applied (Lord et al. 1995a). However, without some form
of control both the diagnostic process and the subsequent
therapeutic endeavors will be corrupted, to the cost of the
patient and to the cost of the health care system.
DISC STIMULATION
Formal studies in normal volunteers have shown that
lumbar disc stimulation is a highly specific test; lumbar
discs do not hurt in asymptomatic individuals (Walsh et aL
1990). Thus, finding a painful disc in a patient is a signi.fi-
cant observation. However, even so, controls are manda-
tory to exclude false-positive responses, i.e., to refute the
competing hypothesis that stimulating any disc reproduces
the patient's pain.
The IASP has recommended that for disc stimulation
to be considered valid, at least one and preferably two ad-
jacent discs be stimulated as controls. For a disc to toe
deemed painful stimulation of that disc but neither of the
adjacent discs should reproduce the patient's pain
(Merskey and Bogduk 1994), For IDD to be diagnosed not

7. only mustthe patient satisfy the criteria for discogenic pain
but also a grade three fissure must be demonstrated onpost
discography CT (Merskey and Bogduk 1994). Unless these
criteria are satisfied, grave doubts can be cast on the diag-
nosis offered.
With respect to cervical discogenic pain, investigators
must be cautious. Cervical disc stimulation is not sup-
ported by literature as robust as that pertaining to the lum-
bar spine (Bogduk et al. 1995a). The cervical discs are not
subject to the same lesions as are lumbar discs, and it is
not convincingly clear that cervical discs do not hurt in
normal volunteers. Moreover, it has been shown that pa-
tients with apparently painful discs can have their pain
relieved by blocks of the zygapophysial joints at the same
segment (Bogduk and Aprill 1993), and that in patients
with cervical myelopathy but no neck pain, disc stimula-
tion is painful in some 50% (Shinomiya et al. 1993).
EPIDEMIOLOGY
In the past, belief in zygapophysialjoint pain, sacroil-
iacjoint pain and discogenic pain has been based on fash-
ion, passion and defiance. Hence, these entities have been
controversial. Recent studies, however, nave brought sci-
ence to'bear.
Using placebo-controlled blocks, the prevalence of
lumbar zygapophysialjoint pain in a sample of Australian,
Rheumatology patients with chronic low back pain was
found to .be 40% with 95% confidence intervals (CI) of
27%-53% (Schwarzer et al. 1995a). Using comparative
blocks in a sample of American patients with back pain,
the prevalence of zygapophysial joint pain was found to be
15% (95% CI: 10%-20%)(Schwarzer et al. 1994a)."
Using comparative local anesthetic blocks, the
prevalence of cervical zygapophysialjoint pain was found
to be 54% in patients with chronic neck pain after whip-
lash (Barnsley et al. 1995), and this figure has nowbeen
confirmed using saline controlled, triple diagnosticblocks
(Lord et aL 1996a). Moreover, headache after whiplash
can be traced in^the majority of patients to a painful C-3
zygapophysialjoint (Lord et at 1994).
Using intra-articular blocks, the prevalence of sacroil-
iacjoint pain'was found to be 13%in patients withchronic
low back pain (95% CI: 6%-20%) (Schwarzer et al.
1995b). This figure will soon be substantiated by a stndy
yet to be published, that employed comparativelocal anes-
thetic blocks.
- The prevalence of cervical discogenic pain has not
been formally studied, but that of lumbar discogenic pain
has. Using the stringent criteria of the IASP, the preva-
lence of IDD amongst patients with chronic lowbackpain
has been found to be 39% (95% CI: 29%-49%)(Schwarzer
et al. 1995c).
With respect to back pain, it has furthermore been
shown that zygapophysial joint pain, IDD and sacroiliac
"joint pain rarely coexist in the same patient (Schwarzer et
aL 1994d). Less than 10% of patients have more than one
condition; the majority suffer either IDD, zygapophysial
joint pain or sacroiliac joint pain, with IDDbeing the most
common entity, accounting for 39% of patients.
These prevalence figures cannot be ignored; they are
based on controlled studies. Contraryto conventional wis-
dom that 70% of low back pain cannot be diagnosed, the
converse is true. With 39% suffering IDD; 15% suffering
zygapophysialjoint pain and 12%suffering sacroiliac joint
pain, a source of pain can be found in over 60% of pa-
tients. Other entities such as dural irritation (Bogduk and
Twomey 1991) and torsion injuries to the disc (Bogduk
1991; Bogduk et al. 1995a) have not been formally studied
but could account for a good proportion ofthe remainder.
These prevalence figures put beyond doubt that pa-
tients with chronic low back pain indeed do suffer' from
identifiable, organic sources of pain. The nihilism and
defeatism commonly applied to low back pain stems from
the failure of most practitioners to implement investiga-
tions that answer the question from where does the pain
stem. If inappropriate tests such as EMG and imaging are
used nothing will be found in the majority of cases, falsely
justifying the impression that nothing can be found, and
that therefore the problem must be a complex bio-psycho-
social one. However, this is not the impression that arises
if and onceprocedures designed for precision diagnosis are
implemented.
In the case ofneck pain, zygapophysialjoint pain has
been shown to be the single most common source of
chronic pain; yet abjectly this is singularly the most over-
looked diagnosis. The reason would appear to be that phy-
sicians are not taught about this diagnosis. Not taught to
consider it, they don't; and the 50% ofpatients who have it
are denied a diagnosis and at best are told they have a bi-
opsychosocial problem, or at worst are accused of malin-
gering or fraud.
A further reason, and one that pertains to all preci-
sion spinal diagnostic procedures is that they reo^iire spe-
cial facilities and special skills; they are not office
procedures. To date, they nave been restricted to special-
ized radiology or spine centers, and have not been em-
braced by leading pain clinics that direct political
correctness in this field. However, this ideological differ-
ence does not refute the scientific propriety of precision
procedures that are executed responsibly under controlled
conditions.
UTILITY
The ultimate measure of any diagnostic procedure is
its therapeutic utility. At the end of the day, does pin-
pointing a diagnosis make any difference to management;
if a source of pain is found, is there a specific treatment
that can completely stop that pain?
For some of the entities addressed above there are
specific treatments; for others, such treatments are still
being developed. To expect a treatment for everycondition
is somewhatunfair and demanding, for the scientific evi-

8. dence of the prevalence of these conditions has not been
available for more than two years in most instances.
Sacroiliacjoint pain is an example of a conditionthat
can be diagnosed but for which there is no established
treatment Conservative measures have not proved effec-
tive, and operative treatments have no decent reputation.
This is clearly a condition that currently rests on the
drawing board.
The only treatment for IDD that makes any sense
anatomically and biomechanically is anterior lumbar inter-
body fusion. Posterior and lateral fusions are not appro-
priate because the offending disc remains in situ and
continues to generate chemical and mechanical nocicep-
tion (Bogduk 1991); persistent discogenic pain has been
demonstrated in patients who have undergone technically
satisfactory posterolateral fusions (Weatherley et al.1986).
In this regard reviews of the literature paint a poor
picture offusion (Turner et al. 1992) but that is because, in
the past, surgeons have applied a variety of fusions indis-
criminately to patients, without formulating a precision
diagnosis, using persistent back pain 'as the sole indication
for operation. In contrast, some surgeons testify to good
results with anterior, lumbar interbody fusion if the proce-
dure is restricted to patients with morphologically abnor-
mal discs on MRJ which prove to be painful upon disc
simulation (Gill—and Blumenthal 1992; Newman and-
Gristead 1992). Such anecdotes, however, do not comple-
ment the science that has been applied to the diagnosis of
IDD. What is critically and urgently required are formal
studies that document the success of fusion and the corre-
lation between success, failure and the response to diag-
nostic disc stimulation. The same applies to cervical
discogenic pain, where the enthusiasm for fusion has not
been matched by scientific diagnostic studies. .Such studies
are currently being fostered by the International Spinal
Injection Society (Bogduk1994b;Derby1994).
An exciting alternative is intra-discal injection of
steroids. In a preliminary audit, it has been found that the
majority of patients with a painful disc that exhibits a
high-intensity zone in its anulus on MRIf steroids offer
good relief of pain to the extent of avoiding planned sur-
gery (Schellhas et al. 1995). It is imperative, however, that
this observation urgently be subjected to a controlled trial
before it is adopted and abused as a false panacea forIDD.
Notwithstanding the optimism that finding a painful
disc may lead to effective therapy, disc stimulation as a
diagnostic test has just as important a negative value. If
disc stimulation fails to reproduce a patient's pain, a diag-
nosis of discogenic pain cannot be rendered. Such a nega-
tive diagnosis has the power to advise against, if not
prevent, gratuitous disc surgery being inappropriately ap-
plied. Thus, disc stimulation has the power to protect pa-
tients from unnecessary surgery as much as, if not more
than, committingthem to surgery.
With respect to lumbar 2ygapophysial joint pain, in-
tra-articular injection therapy has been shown to offer no
lasting benefit; steroids offer no advantage over saline
(Carette et al. 1991). The only other specific therapy that
has been touted is percutaneous radiofrequency (RF) neu-
rotomy. This procedure has a checkered past, largely be-
cause it was fostered by enthusiasm but no scientific
accountability (Bogduk et al. 1995a). In particular, few
proponents ever used anatomically accurate techniques
{Bogduk et al. 1987) Recent studies, however, have shown
a clear relationship between positive response to blocks
and good outcome after treatment; the treatment does not
workin patients who do not respond to blocks (North et al.
1994). This underscores the critical relationship between
precision diagnosis and targeted therapy. However, in the
case of lumbar RF neurotomy, controlled studies have not
yet property vindicated the utility ofthe procedure. This is
not the case for cervical neurotomy.
rntra-articular steroids have been shown to, offer no
worthwhile therapeutic "benefit for cervical zygapophysial
joint pain (Bamsley et al. 1994), but RF neurotonry does.
An open audit of cervical RF neurotorny showed that
treatment of pain from the 'C2-3 zygapophysial joint was
marred by technical problems, but results at levels below
C2-3 were sufficiently impressive to justify a controlled
'trial (Lord et at 1995b). This study has now been com-
pleted (Lord et al 1996b).
Patients were enrolled in the study if they obtained
complete relief of their neck pain following blocks of the
painful joint on two occasions with local anesthetics but
provided they did not obtain relief when challenged with
placebo. All patients underwent a radiofrequency neurot-
omy procedure lasting three hours. The procedure- was
identical in all respects save that half of the patients re-
ceived lesions at 80° C while half received the same elec-
trode placements but no current was delivered. Neither the
patient nor the surgeon knew what was delivered.
Survival analysis revealed that those patients who re-
ceived sham lesions had a median duration of relief of 11
days. The median duration of relief in those who received
active treatment was 227 days at the time of follow-up,
with several patients still continuing with complete relief
of pain.
Cervical RF neurotomyis not a placebo procedure. It
affords complete relief of pain stemming from cervical
zygapophysial joints, in those patients in whom zygapo-
physialjoint pain has been diagnosed by triple, saline con-
trolled, diagnostic blocks.
So complete is this relief that patients score zero or-
not more than 5/100 on a visual analog scale; their total
word score is zero on the McGill Pain Questionnaire; and
they resume normal activities of dairy living. However,
even more striking is that their profiles of psychological
distress, as measured by the SCL-90R, revert to normal
across all scales (Wallis et al. 1996).
DISCUSSION
Proper studies of the treatment of spinal pain are still
in their infancy. The few investigators cornmitted to" this

9. field in recent years have been pre-occupied with the vali-
dation of precision diagnostic techniques; resources have
not been available to pursue studies of therapy.However,
the history and evolution of cervical zygapophysial joint
pain illustrates what can be achieved, and underscoresthe
philosophy ofproponents of spinal injectionprocedures.
Originally, cervical zygapophysialjoint pain was only
a strange belief not credited by Medicine at large (Bogduk
et al. 1988; Bogdukand Aprill 1992), and indeed, formally
questioned (Frymoyer 1989). However, precision diagnos-
tic techniques were developed and shown to have face va-
lidity (Barnsley and Bogduk 1993) and construct validity
(Bamsley et al. 1993b), and to be robust against placebo
(Lord et al. 1995a). Far from being strange, this condition
was shownto be very common (Bamsley et al. 1995) even
under placebo controlled conditions (Lord et al. 1996a). A
specific therapeutic technique was developed and audited
(Lord et al. 1995b) and survived a placebo-controlled trial
(Lordetal. 1996b).
The sequence of systematic studies demonstratesthat
a symptom (neck pain), previously considered intractable
and beyonddiagnosis, can be resolved using a reductionist
approach. The medical model so much maligned in pain
circles, can be successful. Pin-pointing a source of spinal
pain using precision techniques allows therapy to be spe-
cifically targeted and implemented. Moreover, once pain is
completely relieved psyche-social elements of the original
problem melt away.
This has not been the experience with other proce-
dures and other therapies. However, that is because it has
been the fashion to implement new diagnostic procedures
and new therapies without scientific rigor. The responsi-
bility now lies with proponents of precision diagnostic_
techniques to do for disc stimulation, sacroiliac blocks,
lumbar zygapophysialjoint blocks, and other spinalblocks,
what has been achieved for cervical zygapophysial joint
pain—a full cycle of validated diagnostic techniques, epi-
demiologicai studies and proven, targeted therapy. For
lumbar zygapophysial joint pain and IDD, most of this
cycle is complete. The techniques are available; the epi-
demiological studies have been done; the final step is the
link to good therapy. ' '
The glaring feature of precision diagnosis of spinal
pain is that it is politically incorrect. The management of
spinalpain is currently attended by defeatism and nihilism.
It seems to have been accepted that 70% of spinal pain
cannotbe diagnosed; yet the converse is true.
Precision diagnostic procedures legitimately answer
the crucial question: From where does the pain come?
They do not rely on pattern recognition. Instead, investiga-
tors follow this algorithm:
1. the patient complains of pain.
2. from where could this pain arise?
3.
4.
5.
6.
in the neck the pre-test probability of zygapophysial
joint pain is 50%; therefore, test for zygapophysial
joint pain.
in the back, the pre-test probability of lumbar zyga-
pophysialjoint pain is 15%to 40%; therefore, test for
lumbar zygapophysialjoint pain,
the pre-test probability of sacroiliac joint pain is 12%;
therefore, if zygapophysial joint blocks are negative,
test for sacroiliac joint pain,
the pre-test probability of lumbar IDD is 39%; there-
fore, if synovia! joint blocks are negative, test for
IDD:
Such an algorithm clarifies 50% of neck pain, and
60% of back pain—a substantial 'improvement on minus
70%.
Critics of precision spinal diagnosis should re-
appraise their stance in view of the literature presented
here. Opposition should not be based on the poor reputa-
tion of these procedures because they were performed in-
discriminately in the past or without controls; that should
no longer be an acceptable standard of we. Nor should op-
position arise because the required techniques are not
available in one's local clinical environment. If physicians
wish to make precision diagnoses they should engineer to
have appropriatetechniques made available. Not to do so
means that the benefits of scientific progress in this field
are being denied to patients for cultural and political rea-
sons.
REFERENCES
Aprill C,Bogduk N. High intensity zone: a pathognomonic sign
of painful lumbar disc on MRI, Br J Radiol 1992: 65:361-
369.
Aprill C,Dwyer A, Bogduk N. Cervical zygapophysial joint pain
patterns H: a clinical evaluation. Spine 1990;15:458-461.
Bamsley L, Bogduk N. Medial branch blocks are specific for the
diagnosis of cervical zygapophysial joint pain. Regional
Anaesthesia 1993;18:343-350.
Bamsley L, LordS, Wallis B, BogdukN.False positive ratesof
cervical zygapophysial joint blocks. Clin J Pain
1993a;9:124-130.
Bamsley L, Lord S, BogdukN. Comparative anaesthetic blocks
in the diagnosis of cervical zygapophysial joints pain. Pain
1993b;55:99-106.
Bamsley L, Lord SM, Wallis BJ, Bogduk N. Lack of effect of
intraarticular corticosteroids for chronic pain, in the cervical
zygapophysealjoints. N Engl J Med 1994;330:1047-1050.
BarnsIeyT., Lord SM, WallisBJ, BogdukN7TnVprevalence"6
chronic cervical zygapophysial joint pain after whiplash.
_
BogdukNTTnTcGScaTanatomy ofthe cervical dorsal rami.
Spine 1982;7:319-330.
Bogduk N: The lumbar disc and low back pain, Neurosurg Clin
North Am 19912:791-806.

10. BogdukN. Sources oflow back pain. Iru Jayson MIV (Ed) The
Lumbar Spine and Back Pain. Edinburgh: ChurchillLiv- .
ingstone, 1992, pp 61-88.
BogdukN. Diskography. Am Pain Soc J 1994;3:149-154.
Bogduk B: Proposed discography standards. ISIS Newsletter,
Volume 2, Number 1 .Daly City, CA: International Spinal
Injection Society, 1994, pp 10-13.
Bogduk N. Neck pain, assessment and managementingeneral
practice. Mod Med 1995;38:102-108.
Bogduk N, Aprill C. The prevalence of cervical zygapophysial
joint pain: a first approximation. Spine 1992;17:744-747.
Bogduk N. Aprill C. On the nature of neck pain, discography and
cervical zygapophysial joint blocks. Pain 1993;54:213-217.
Bogduk N, Marsland A The cervical zygapophysialjoints asa
source ofneck pain. Spine 1988;13:610-617.
Bogduk N, TwomeyL. Clinical Anatomy of the LumbarSpine,
2nd ed. Melbourne: Churchill Livingstone, 1991.
Bogduk N, TynanW, Wilson AS. The nerve supplyto the human
lumbar intervertebral discs. J Anat 1981;132:39-56,
Bogduk N, Wilson AS, Tynan W. The humanlumbar dorsal
rami. JAnat 1982;134:383-397.
BogdukN, Macintosh JE, Marsland A:A technical causefor
clinical failure with percutaneous radiofrequency periph-
eral nenrdtorny. Neurosurgery 1 987^0:529-535.
Bogduk N, Windsor M, Inglis A. The innervation ofthe cervical
intervertebral discs. Spine 1988;13:2-8.
Bogduk N,-Aprill C, Derby R. Discography. In: WhiteAH (Ed),
Spine Care, Volume One: Diagnosis andConservative
Treatment. St Louis: Mosby, 1995a, pp 219-238.
Bogduk N, Aprill C, Derby R. Diagnostic blocks of synovia!
joints. In: Wnite AH (Ed), Spine Care, VolumeOne:Diag-
nosis and Conservative Treatment. St Louis: Mosby,
1995b,pp 298-321.
Bogduk N, Aprill C, Derby R. Lumbarzygapophysialjoint pain:
diagnostic blocks and therapy. In: Wilson DJ (Ed), Inter-
ventional Radiology of the Musculoskeletal System.Lon-
don: EdwardArnold, 1995c, pp 73-86.
Butterworth JF, Strichartz GR. Molecular mechanisms of local
anesthesia: a review. Anesthesiology 1990;72:711-734.
Carette S,Marcoux S, Tnichon R, et al. A controlled trial of
corticosteroid injections into facet joints for chronic low
back pain. NEnglJ Med 1991;325:1002-1007.
Ch.ua WH, BogdukN. The surgical anatomyof thoracic facet
denervation, Acta Neurochir 1995;136:140-144.
Derby R:A secondproposal for discographystandards. ISIS
Newsletter, Volume 2, Number 2. Daly City, CA:Interna-
tional Spinal Injection Society, 1994, pp 108-122.
Deyo RA, Diehl AK, Lumbar spine films in primarycare:current
use and effects of selective ordering criteria, J GenInt Med
1986;l:20-25.
Dreyfuss P, Tibiletti C, Dreyer SJ. Thoracic zygapophysealjoint
pain patterns. A study in normal volunteers. Spine
1994a;19:807-811.
Dreyfuss P, Michaelsen M, Fletcher D. Atlanto-occipital and
lateral atlanto-axial joint pain patterns. Spine
1994b;19:1125-1131.
Dwyer A, Aprill C, BogdukN. Cervical zygapophysialjoint pain
patterns L a study in normal volunteers. Spine
Fortin JD, DwyerAP, West S, Pier J: Sacroiliacjoint pain refer-
ral mapsupon applying a new injection/arthrography tech-
nique: part t asymptomatic volunteers. Spine
1994;19:1475-1482.
Friedenberg ZB, Miller WT. Degenerative disk disease of the
cervical spine. J Bone Joint Surg 1963;45A:1171-1178.
Frymoyer JW. Commentary on the cervical zygapophysialjoints
as a source ofneck pain. Yearbook of Orthopedics, 1989,
pp278-279.
Gill K, Blumenthal SL:Functional results after anteriorlumbar
fusion at L5-S1 in patients with normal and abnormalMRI
. scans. Spine 1992;17:940-942.
Grob KR, Neuhuber WL, KJssling RO. mnervation ofthe human
sacroiliacjoint Z Rheumatol 1995;54:117-122.
Green GJ, Baljet B, Drukker J: Nerves and nerve plexuses of the
human vertebral column. Am JAnat 1990;188:282-296.
J6nsson H, Jr., Bring G, Rauschning W, Sahlstedt B: Hidden
cervical spine injuries in traffic accident victims with skull
fractures. J SpinDis 1991;4:251-263.
Lawrence JS, Bremner JM, Bier F: Osteoarthrosis. Prevalence in
the populationand relationship between symptomsand X-
ray changes.Am RJieurn Dis 1966;25:l-24.
Lord SM, Barnsley L, BogdukN. Percutaneous radiofrequency
neurotomy in the treatment of cervical zygapophysialjoint
pain: a caution. Neurosurgery 1995;36:732-739.'
Lord SM,Barnsley L, BogdukN. The utility of comparativelocal
anaesthetic blocks versus placebo-controlled blocks for the
diagnosis ofcervical zygapophysialjoint pain. Clin JPain
1995;208-213.
Lord S,Barnsley L, Wallis B, BogdukN. Third occipital head-
ache: a prevalence study. J Neurol Neurosurg Psychiat
1994;57:1187-1190.
Lord S, Barnsley L, Wallis BJ, Bogduk N. Chronic cervical zyga-
pophysialjoint pain after whiplash: a placebo-controlled
prevalence study. Spine 1996a; (in press)
Lord S, Barnsley L, Wallis BJ, Bogduk N. A randomised, double-
blind, controlledtrial ofpercutaneous radiofrequency neu-
rotomy for the treatment of chronic cervicalzygapophysial
joint pain. N Engl J.Med 1996b (submitted)
Magora A, Schwartz TA: Relation between the low back pain
syndrome and X-ray findings. Scand J Rehab Med
1976;8:115-125.
McCall IW, Park WM, O'Brien JP: Induced pain referral from
posterior lumbar elements in normal subjects. Spine
1979;4:441-446.
Mendel T, Wink CS, Zimny ML. Neural elements inhuman
cervical intervertebral discs. Spine 1992;17:132-135.
Merskey H, Bogduk N (Eds). Classification of Chronic Pain.
Descriptions of Chronic Pain Syndromesand Definition of
Pain Terms, 2nd ed. IASP Press, Seattle, 1994.
Moneta GB,Videman T, Kaivanto K, et al. Reported pain during
lumbar discographyas a function of anular rupturesand
disc degeneration. A re-analysis of 833 discograms. Spine
1994;17:1968-1974.
Mooney V, Robertson J: The facet syndrome. Clin Orthop
1976;115:149-156.
Newman MH, Gristead GL:Anterior lumbar interbody fusion for
internal disc disruption. Spine 1992;17:831-833.
North RE, Han M, Zahurak M, Kidd DH: Radiofrequency lumbar
facet denervation: analysis of prognostic factors. Pain
1994;57:77-83.
Schellhas KP, Pollei SR, Dorwart RH, Thoracic discography.A
safe andreliable technique. Spine 1994;19:2103-2109.
Schellhas KP. HeithofTKB, Pollei SR, Lumbar disc high inten-
sity zone:managementwith intradiscal steroids and local
anaesthetics. Proceedings of the 3rd AnnualScientific
Meeting ofthe International Spinal Injection Society,New

11. Orleans, September 23-24,1995. Tntematioiifll SpinalIn-
jection Society: San Francisco, 1995, p 103.
Schwarzer AC, Aprill CN, Derby R, et al Clinical features of
patients with pain, stemming from the lumbar zygapo-
physialjoints. Is the lumbar facet syndrome a clinical en-
tity? Spine 1994a;19:1132-1137.
Schwarzer AC, Derby R, Aprill CN, et al. Painfrom,the lumbar
zygapophysialjoints: atestoftwo models.-J Spinal Disord
1994b;7:331-336.
Schwarzer AC, Aprill CN, Derby R, et al. The false-positive rate
of uncontrolled diagnostic blocks ofthe lumbar zygapo-
physial joints. Pain 1994c;58:195-200. •.
Schwarzer AC, Aprill CN, Derby R, et al. The relative contribu-
tions ofthe disc and zygapophysealjoint in chronic low
• backpain. Spine 1994d;19:801-806.
Schwarzer AC, Wang S, Bogduk N, McNaught PJ, Lament R.
Prevalence and clinical features of lumbar Zygapophysial
joint pain: a study in an Australian population with chronic
low backpain. Am Rheum Dis 1995a;54:100-106.
Schwarzer AC, Aprill CN, Bogduk N. The sacroiliacjoint in
chronic low back pain. Spine 1995b;20:31-37.-
Schwarzer AC, Aprill CN, Derby R, et al. The prevalence and
clinical features of internal disc disruption in patients with
chronic lowback pain. Spine 1995c£0:187S-1883,1995.
Schwarzer AC, Wang S, O'Driscoll D, et aL The ability of com-
.puted tomography to identify a painful zygapophysialjoint
in patients with chronic low back pain. Spine
1995d;20:907-912.
Shinomiya YNakao K, Shindoh S, Mochida K, Furuya K.
Evaluation of cervical discography in pain origin and
provocation. J Spinal Dis 1993;6:422^26.
Taylor JR,Tvomey LT, Corker M: Bone and soft tissue injuries
inpost-mortem lumbar spines. Paraplegia 1990;28:119-
129.
Twomey LT, Taylor JR, Taylor MM: Unsuspected damageto
lumbar zygapophyseal (facet) joints after motor vehicle ac-
cidents. Med J Aust 1989;151:210-217.
Turner JA, Ersek M, Herron L, et al. Patient outcomes after lum-
bar spinal fusions. JAMA 1992^268:907-911.
Vanharanta H, Sachs BL, Spivey MA, et al. The relationship of
pain provocation to lumbar disc deterioration as seenby
CT/discography. Spine 1987;12:295-298,
Wallis BJ, Lord S, Bamsley L, BogdukN, Changes in the psy-
chological profile of whiplash patients following treatment
with radiofrequency neurotomy: a randomised, double-
blind, placebo-controlled trial. Pain 1996 (submitted).
Walsh TR, Weiastein JN, Spratt KF et al. Lumbar discography in
normal subjects. JBone Joint Surg 1990;72A;10S1-1088.
Weatherley CR, Prickett CF, O'Brien JP. Discogenic pain per-
sisting despite solid posterior fusion. J BoneJoint Surg
1986;68:142-143.
Correspondence to: Professor N. Bogduk, Faculty of
Medicine and Health. Sciences, University of Newcastle,
University Dr, Callaghan, NSW 2308, Australia. Tel: 61-
49-215608; Fax: 61-49-216903.
323