The Spine Journal 4 (2004) 2735

2003 Outstanding Paper Award: Nonoperative Science

A randomized, placebo-controlled trial of intradiscal electrothermal
therapy for the treatment of discogenic low back pain
Kevin J. Pauza, MDa,b,*, Susan Howell, BS, RNb, Paul Dreyfuss, MDb, John H. Peloza, MDc,
Kathryn Dawson, PhDd, Nikolai Bogduk, MD, DSce
a
Texas Spine and Joint Hospital, 1814 Roseland Boulevard, Tyler, TX 75701, USA
b
East Texas Medical Center Neurological Institute, Tyler, TX 75701, USA
c
Center for Spine Care, Dallas, TX 75231, USA
d
Virginia Commonwealth University, Richmond, VA 23219, USA
e
University of Newcastle, Royal Newcastle Hospital, Newcastle, NSW 2300, Australia
Received 3 February 2003; accepted 2 July 2003

Abstract BACKGROUND CONTEXT: Intradiscal electrothermal therapy (IDET) is a treatment for disco-
genic low back pain the efficacy of which has not been rigorously tested.
PURPOSE: To compare the efficacy of IDET with that of a placebo treatment.
STUDY DESIGN/SETTING: Randomized, placebo-controlled, prospective trial.
PATIENT SAMPLE: Patients were recruited by referral and the media. No inducements were pro-
vided to any patient in order to have them participate. Of 1,360 individuals who were prepared
to submit to randomization, 260 were found potentially eligible after clinical examination and 64
became eligible after discography. All had discogenic low back pain lasting longer than 6 months,
with no comorbidity. Thirty-seven were allocated to IDET and 27 to sham treatment. Both groups were
satisfactorily matched for demographic and clinical features.
METHODS: IDET was performed using a standard protocol, in which the posterior annulus of
the painful disc was heated to 90 C. Sham therapy consisted of introducing a needle onto the disc
and exposing the patient to the same visual and auditory environment as for a real procedure. Thirty-
two (85%) of the patients randomized to the IDET group and 24 (89%) of those assigned to the
sham group complied fully with the protocol of the study, and complete follow-up data are available
for all of these patients.
OUTCOME MEASURES: The principal outcome measures were pain and disability, assessed
using a visual analog scale for pain, the Short Form (SF)-36, and the Oswestry disability scale.
RESULTS: Patients in both groups exhibited improvements, but mean improvements in pain,
disability and depression were significantly greater in the group treated with IDET. More patients
deteriorated when subjected to sham treatment, whereas a greater proportion showed improvements in
pain when treated with IDET. The number needed to treat, to achieve 75% relief of pain, was
five. Whereas approximately 40% of the patients achieved greater than 50% relief of their pain,
approximately 50% of the patients experienced no appreciable benefit.
CONCLUSIONS: Nonspecific factors associated with the procedure account for a proportion of
the apparent efficacy of IDET, but its efficacy cannot be attributed wholly to a placebo effect. The
results of this trial cannot be generalized to patients who do not fit the strict inclusion criteria

FDA device/drug status: approved for this indication: IDET Spine Cath. * Corresponding author. Texas Spine and Joint Hospital, 1814 Roseland
Coinvestigator KD acknowledges a financial relationship (consultant Boulevard, Tyler, TX 75701, USA. Tel.: (903) 593-2222; fax: (903) 593-
for Smith and Nephew Endoscopy, Andover, MA) that may indirectly relate 0142.
to the subject of this research. E-mail address: kevinpauza@tyler.net (K.J. Pauza)

1529-9430/04/$ see front matter 2004 Elsevier Inc. All rights reserved.
doi:10.1016/j.spinee.2003.07.001
28 K.J. Pauza et al. / The Spine Journal 4 (2004) 2735
of this study, but IDET appears to provide worthwhile relief in a small proportion of
strictly defined patients undergoing this treatment for intractable low back pain. 2004 Elsevier
Inc. All rights reserved.
Keywords: Back pain; Disc treatment; Intradiscal electrothermal therapy; IDET; Controlled trial; Placebo; Discogenic pain
Introduction
Intradiscal electrothermal therapy (IDET) was developed
in 1997 as a treatment for lumbar discogenic pain [1]. It
involves inserting a flexible electrode in a circumferential
fashion into the annulus fibrosus of the painful disc (Fig.
1). Heating the electrode is believed to relieve pain by mech-
anisms that have not been established. Conjectures include
that heating the electrode coagulates nerve fibers in the
annulus [1,2], stiffens collagen [1], seals fissures in the
annulus [2] or coagulates inflammatory chemicals in those
fissures [2].
The procedure was promoted on the basis of observational
studies, initially in the form of conference papers [35]
but subsequently in the form of refereed papers [610]. A
cohort study, using a convenience sample of patients whose
insurers denied treatment with IDET, reported that IDET
achieved better outcomes than a rehabilitation program
[11,12].
Fig. 1. Electrode insertion for intradiscal electrothermal therapy. (Top)
Posteroanterior view. (Bottom) Lateral view.
Nevertheless, the promotion of IDET has been contested
[1317], because no studies have compared IDET with any
other treatment in a randomized controlled trial; nor has a
placebo effect been tested. The present study was initiated
to redress this deficiency.
Methods
The study was undertaken in accordance with both the
prescribed guidelines for randomized controlled trials [18]
and the quality checklists of systematic reviews for treat-
ments for low back pain [1922]. It was conducted in a
small city in Texas, from a private practice that specialized in
the diagnosis and treatment of spinal pain. The institutional
review board of the East Texas Medical Center approved
the study. The objective was to test the efficacy of IDET
against a placebo intervention.
The study was designed to unblind patients at 6 months,
because it was believed unethical to subject patients who
did not obtain relief to a longer duration of untreated low
back pain. Asking patients to remain untreated risked their
noncompliance with follow-up evaluation and increased
their likelihood of resorting to treatments outside the investi-
gations protocol. Furthermore, based on clinical experience
at the time, it was expected that outcomes at 6 months
would be stable, and any differences discerned after a longer
duration would be nominal. This contention was validated
by the results of another study that demonstrated that the
outcomes at 6 months after IDET are stable, and they
are predictive of outcomes at 12 months and 24 months
[12]. Patients who by 6 months failed to obtain relief from
sham treatment were to be offered IDET at that time.
Between September 2000 and April 2002, patients for the
study were recruited from the practice of the senior author,
from colleagues and by invitations in local and national
television and print media. No inducements were provided to
any patient in order to have them participate. Interested
patients were initially screened by a telephone interview
to assess potential eligibility using a 21-question survey.
Patients who appeared eligible attended a face-to-face inter-
view and physical examination to confirm eligibility. Pri-
mary eligibility criteria were as follows: age 18 to 65 years;
low back greater than leg pain present for greater than 6
months duration; failure to improve after at least 6 weeks
of nonoperative care, including anti-inflammatory and anal-
gesic medications and a physical therapy and/or home di-
rected lumbar exercise program; low back pain exacerbated
 K.J. Pauza et al. / The Spine Journal 4 (2004) 2735
by sitting or standing and relieved by lying down; a score
less than 20 on the Beck depression scale; no surgical inter-
ventions within the previous 3 months and less than 20%
disc height narrowing on lateral plain film radiographs.
The exclusion criteria were as follows: previous lumbar
spine surgery; abnormal neurological examination other than
ankle reflex changes; radicular pain by history or examina-
tion; structural deformities, such as spondylolisthesis at the
painful segmental level, vertebral canal stenosis or scoliosis;
intervertebral disc herniations greater than 4 mm; sequest-
ered intervertebral disc herniations; concomitant cervical or
thoracic pain greater than 2/10 on a VAS; uncontrolled
or acute medical illnesses; chronic severe conditions, such
as rheumatoid arthritis; ambulatory dysfunction; pregnancy;
workmans compensation; injury litigation; disability remu-
neration; allergy to contrast media or drugs to be used in the
intended procedure and unwillingness to consent to the study.
Patients considered potentially eligible underwent pro-
vocation discography to test if they had discogenic pain.
Discography was performed according to the standard pre-
scribed by the International Association for the Study of
Pain [23] and the International Spinal Injection Society
[24]. Patients were positioned prone with minimal or no
conscious sedation. Individual 22-gauge spinal needles were
inserted into the nucleus pulposus of the caudal three or four
lumbar intervertebral discs. Radiopaque contrast medium
(iopanidol 61% with iodine 30%) with antibiotic (cephazolin
0.056 gm/cc or gatifloxacin 0.33 mg/cc) was injected, and
using digital pressure manometers (Merit Medical Systems,
South Jordan, UT) [25], the opening pressure and the pressure
at which pain provocation first occurred were recorded along
with volumes of injection. Each disc was pressurized on at
least two separate occasions without the patient knowing
which level was being tested. Several randomly interspersed
sham pressurizations were performed. A patient was deemed to
have discogenic pain if provocation of an intervertebral disc
reproduced their accustomed pain within similar pressure
ranges but provided that no pain was reported when adjacent
discs were stimulated or during sham pressurizations.
Computed tomography (CT) scans were obtained prompt-
ly after discography to assess the internal architecture of
the painful disc. To be eligible for the study, patients had to
have a posterior tear of the annulus fibrosus. Patients with
only anterior or lateral tears or with diffuse changes in the
disc were excluded. All consecutive patients who satisfied
the criteria for discogenic pain in a disc with a posterior
tear were invited to participate in the trial.
Using computer-generated, random numbers, all consecu-
tive patients who agreed to participate were randomized to
either IDET or a sham treatment. Before initiating treatment,
all patients received an intravenous infusion of prophylactic
antibiotics. All were informed that a needle would be placed
in their back, which would be displayed on the fluoroscopic
monitor. Meanwhile, they would all perceive sensations in
their lumbar spine and hear sounds from the equipment that
would not reflect whether they received the treatment or the
29
sham. Only after the patient received conscious sedation and
a 17-gauge introducer needle had been inserted to contact
the outer annulus fibrosis, the randomization schedule was
revealed, for the first time, to the principal investigator. The
procedure was then completed according to which treatment
the patient had been allocated.
The active treatment involved first advancing the pre-
viously placed introducer needle into the target disc. Thereaf-
ter, the flexible electrode [26] was passed through the needle
into the disc and navigated until it assumed a circumferential
placement within the annulus fibrosus or along the interface
between the nucleus pulposus and annulus fibrosus (Fig.
1). Complete coverage of the posterior circumference was
confirmed by lateral radiograph demonstrating that the proxi-
mal and distal radio-opaque markers on the catheter were
radiographically equidistant in the anteroposterior direction.
Once satisfactory placement was obtained, the electrode was
heated to a temperature of 90 C, following a standard proto-
col [1]. If pain was produced during heating, 50 to 100 g
of fentanyl was administered. Once the heating had been
completed, the electrode was withdrawn and 1 cc of bupiva-
caine 0.75% mixed with antibiotic (cephazolin 0.056 gm/cc
or gatifloxacin 0.33 mg/cc) was injected into the disc.
The sham treatment commenced with the introducer
needle firmly positioned against the outer most aspect of
the annulus fibrosis. Instead of introducing the electrode, the
patient was exposed to a fluoroscope monitor demonstrating
the passage of an electrode. After a time period elapsed
equal to that necessary for electrode placement, generator
noises were produced for 16.5 minutes to mimic an active
treatment.
After treatment, the patient was never in contact with
anyone privy to the randomization assignment. Subjects in
both groups were prescribed hydrocodone 10 mg/acetamino-
phen 325 mg for any postprocedural pain. Both groups un-
derwent a monitored postoperative rehabilitation program
that consisted of wearing a lumbar corset for 6 weeks and
a supervised lumbar spine stabilization exercise program
beginning at Week 6 and progressing into an independent
exercise program by Week 12. Exercise compliance was
assessed by a physical therapist and a physician blinded to
their randomization assignment.
Outcomes were assessed using a battery of instruments.
These were a 10-point visual analog scale (VAS) for pain
[2730], the Short Form (SF)-36 [3133] and the Oswestry
Disability Scale [34]. Scores on these instruments were ob-
tained before treatment and at 6 months after treatment by
an evaluator who was blinded to which treatment the pa-
tient received.
For continuous variables, group means and standard devi-
ations were determined and compared using a t test. Fre-
quency distributions and Fishers exact test were used to
summarize and compare categorical variables. The primary
objective of the study was to compare the improvement
in pain and physical function between groups. Secondary
objectives included reporting the adverse event profile of
30 K.J. Pauza et al. / The Spine Journal 4 (2004) 2735
IDET. Post hoc analyses were also conducted that summa-
rized the proportions of patients who at 6 months obtained
25%, 50% and 75% relief of pain relative to pretreatment
levels. Percentage relief of pain at follow-up was calculated,
for each patient, as the difference between their pain score
from baseline, divided by their baseline score, and converted
to a percentage. The number needed to treat [35] was calcu-
lated for achieving 75% relief of pain. A post hoc analysis
was also conducted to determine the efficacy of IDET and
of sham treatment for patients with different pain scores,
and different degrees of disability, at inception.
In order to have 80% power to detect a difference as
small as 2.0, for a standard deviation of 2.8, and using an
allocation of 3:2 for active and control treatment, the study
required at least 40 patients to be treated with IDET and 27
to be treated with sham therapy. In principle, having recruited
fewer patients than originally planned reduced the power of
the present study from 80% to 60%. In practice, however,
low power is relevant only when studies fail to detect a
significant difference. Failure to detect a difference would
constitute a Type 2 error. However, this does not apply in
the present study, for differences were, indeed, detected.
It transpired that 60% power was enough to detect the
significant differences reported.
Fig. 2. Profile of a randomized, placebo-controlled trial of intradiscal electrothermal therapy (IDET) for chronic low back pain.
Results
Publicizing the study attracted enquiries from 4,253 pa-
tients. Of these, 1,360 were considered potentially eligible
after the telephone interview and underwent consultation and
physical examination. The majority was determined to be
ineligible because they had excessive loss of disc height,
radicular pain, were unable to travel or were unwilling to
consent to randomization or frequent evaluations. Of these
1,360 patients, 260 were confirmed as eligible and were will-
ing to consent to the trial and undergo discography. These
260 patients underwent discography, and 64 met criteria and
enrolled. After randomization, 37 were assigned to undergo
IDET and 27 to sham treatment (Fig. 2).
With respect to demographic and clinical features, there
were no relevant statistically significant differences between
the two groups at inception (Table 1). Nor were there any
significant differences in baseline values of the outcome
measures (Table 2). Although less than normal, scores for
most of the subscales of the SF-36 were high, indicating that
the patients in both groups were not particularly disabled,
on average, in terms of general health, mental health, role
emotional or social functioning. Nor were they depressed.
In essence, the patients were reasonably healthy apart from
having pain and slight to moderate disability in physical
 K.J. Pauza et al. / The Spine Journal 4 (2004) 2735 31

Table 1 Table 2
Full analysis of demographic and clinical features of patients assigned Full analysis of baseline outcome measures of patients assigned to
to undergo intradiscal electrothermal therapy or sham treatment undergo intradiscal electrothermal therapy or sham treatment
IDET Sham IDET Sham
(n37), (n27), (n37), (n27),
Feature n % n % p Value* Outcome measure Mean SD Mean SD p Value*
Male 19 51 11 41 .454 VAS for pain (010) 6.5 1.6 6.5 1.8 .994
Female 18 49 16 59 SF-36 (0100)
Age (meanSD), 4210 408 .446 Bodily Pain 35 12 35 13 .866
years Physical Functioning 54 24 48 21 .244
Work status: .251 Role Physical 39 40 16 24 .010
Unemployed 3 8 4 15 Role Emotional 84 32 84 28 .983
because Social Functioning 72 24 69 24 .684
of back pain Mental Health 79 15 78 13 .839
Unemployed not 3 8 2 7 General Health 73 18 72 17 .731
because of Vitality 51 20 47 22 .396
back pain Oswestry Disability 32 10 33 11 .632
Homemaker 1 3 4 15 Scale (0100)
Working 30 81 17 63 * The p values pertain to t tests.
Disability payments .712 IDETintradiscal electrothermal therapy; SFShort Form; VASvi-
Yes 3 11 4 15 sual analog scale.
No 34 89 23 85
Work class .376
Manual 14 38 8 30
Sedentary 10 27 12 44
Mixed 13 35 7 26 satisfactory placement of the electrode during treatment
Duration of pain .749 and was excluded immediately afterward. One had a lower-
612 months 2 5 3 11 extremity fracture, shortly after undergoing treatment, which
1224 months 6 16 4 15
24 months 29 78 20 74
impaired ambulation and, therefore, the assessment of his
Previous treatment response. Two subjects had new injuries not related to their
Physiotherapy 15 41 19 70 .024 initial diagnosis. An additional two subjects revealed pre-
Bed rest 16 43 15 56 .448 viously undisclosed exclusion criteria during follow-up. Of
Exercise 18 49 17 63 .314
Anti-inflammatory 29 78 21 78 1.000 those two, one had an undisclosed concurrent neurological
drugs illness and the other had an undisclosed compensation claim,
Opioids 23 62 14 52 .451 while covertly obtaining opioids from other sources. The
Injections 22 60 19 70 .436
per-protocol analysis of the outcome data captured 32 (86%)
Back pain 37 100 27 100 1.000
Referred pain of the patients randomized to the IDET group and 24 (89%) of
In buttock 12 32 4 15 .147 the sham group. None of these patients resorted to any
In thigh 13 35 9 33 1.000 cointervention or self-medication outside the postoperative
In leg 9 24 8 30 .776
Discs painful and .621
analgesics and rehabilitation program that were prescribed.
treated No patient had any adverse effects attributable to their
L5S1 14 38 7 26 treatment.
L4L5 7 19 8 30 Blinding was satisfactory. Within 1 hour after treatment,
L3L4 1 3 3 11
L4L5, L5S1 12 32 8 30 patients were queried as to what treatment they believed
L3L4, L4L5 1 3 0 0 they received, and of those who had IDET, 29 (78%)
L3L4, L5S1 1 3 1 4 believed that they had the active treatment, 2 (5%) believed
L2L3, L4L5 1 3 0 0
they had the sham treatment and 6 (16%) were unable to tell.
* The p values pertain to t tests for continuous data and Fisher exact Of those who had the sham treatment, 21 (74%) believed
tests for categorical data.
IDETintradiscal electrothermal therapy.
they had the active treatment, 2 (7%) believed they had the
sham treatment and 4 (14%) could not tell.
Scores for General Health, Mental Health, Role Emo-
functioning, as seen both on the SF-36 and on the Oswes- tional, and Social Functioning on the SF-36 were quite high
try scale. at baseline and improved only slightly (by less than 10%)
After treatment, eight patients violated the prescribed in both groups and with no statistically significant differences
protocol mandating their rejection from analysis. Of those between groups. Scores for Role Physical improved in both
eight, five were from the IDET group and three from the sham groups, but there was no statistically significant difference
group (Fig. 2). One subject died from unrelated causes. One in the magnitude of the improvement between groups when
subject was lost to follow-up. Another subject did not have a adjusted for the difference in this variable seen at baseline.
32 K.J. Pauza et al. / The Spine Journal 4 (2004) 2735

Both groups exhibited significant improvements in pain Table 4

scores, but the improvement in the IDET group was signifi- Per protocol analysis of absolute and relative changes in pain
scores for patients treated with intradiscal electrothermal therapy or
cantly greater than in the sham group (p.045; Table 3). A sham treatment
similar pattern of improvement occurred in the Bodily Pain
IDET
scores of the SF-36 but was not significantly different (n32), Sham
between the two groups (p.086). Both groups improved Outcome n % (n2), % p Value
with respect to Physical Functioning on the SF-36 (p.050), Pain (010) .037
although with no significant difference between groups; but worse 2 6 8 33
on the Oswestry Disability Scale, the IDET group achieved same 5 16 5 21
significantly better outcomes (p.050). improvement 2.0 7 22 2 8
With respect to categorical outcomes, statistically sig- improvement 2.0 18 56 9 38
Change pain (%) .028
nificant differences in favor of the IDET group occurred 0 2 6 8 33
both for absolute change and for relative changes in pain 024 11 34 6 23
scores as measured by the VAS (Table 4). Much of this 2549 7 22 2 8
difference could be attributed to a greater proportion of 5074 5 16 7 29
patients who underwent sham treatment having deteriorated, 7599 4 13 0 0
100 3 9 1 4
but also the proportions of patients who achieved greater than
The p values pertain to Fishers Exact Test. The percentage change in
2 points improvement in pain and who achieved greater
pain was calculated as (Initial pain score-Final pain score) 100%.
than 75% relief of pain were greater in the IDET group. IDETintradiscal electrothermal therapy.
Indeed, for achieving greater than 75% relief of pain, the
number needed to treat with IDET is 5.
More substantial differences were found when the main
Discussion
outcomes were stratified according to baseline scores. It
emerged that IDET was significantly more effective for pa- In the social setting in which the present study was
tients with pain scores less than 70 at inception and for conducted, many factors militate against conducting a
patients with poor function or greater disability at inception
(Table 5). Reciprocally, IDET had no significantly greater
effect than sham treatment for patients with low disability Table 5
or who already had good physical function. Per protocol analysis of patients who underwent intradiscal electrothermal
therapy or sham treatment, stratified by baseline scores
IDET (n32) Sham (n24)
Outcome measure n Mean SD n Mean SD p Value*
Table 3 VAS for pain (010)
Main outcomes of patients who underwent intradiscal electrothermal Pain 7 before 22 7.4 0.6 11 8.1 1.0 .023
treatment or sham treatment treatment
IDET (n32) Sham (n24) 6 months 4.9 2.5 5.9 2.7 .290
Change 2.5 2.3 2.1 2.8 .695
Outcome measure Mean SD Mean SD p Value* Pain 7 before 10 5.0 1.3 13 5.2 1.3 .684
VAS for pain (010) treatment
Pretreatment 6.6 1.4 6.5 1.9 .758 6 months 2.6 2.3 5.0 2.6 .032
6 months 4.2 2.6 5.4 2.7 .089 Change 2.3 2.4 0.2 2.7 .039
Change 2.4 2.3 1.1 2.6 .045 SF-36: Physical Functioning (0100)
Function 55 23 69 10 10 68 8 .833
SF-36: Bodily Pain (0100) before treatment
Pretreatment 36 12 35 12 .765 6 months 73 21 76 11 .659
6 months 53 19 44 20 .085 Change 4 21 8 11 .585
Change 17 19 9 15 .086 Function 55 9 23 15 35 15 .087
before treatement
SF-36: Physical Functioning (0100) 6 months 66 25 49 24 .101
Pretreatment 56 24 49 21 .236 Change 43 19 14 20 .002
6 months 71 22 60 24 .079 Oswestry Disability Scale (0100)
Change 15 27 11 17 .548 Disability 40 27 28 7 16 27 8 .815
before treatment
Oswestry Disability Scale (0100) 6 months 18 11 22 13 .333
Pretreatment 31 10 33 11 .485 Change 9 10 5 14 .268
6 months 20 12 28 15 .023 Disability 40 5 50 3 8 45 4 .032
Change 11 11 4 12 .050 before treatment
* The p values pertain to t tests. 6 months 30 13 42 7 .062
IDETintradiscal electrothermal therapy; SFshort form; VAS Change 20 13 4 8 .017
visual analog scale. * The p values pertain to t tests.
 K.J. Pauza et al. / The Spine Journal 4 (2004) 2735
randomized and placebo-controlled trial. In the present con-
text, IDET was readily available from other sources in
conventional practice. This limited the population from
which the present sample could be drawn. Patients did
not need to participate in a trial if they could obtain the
treatment elsewhere. Others have reported similar problems
when attempting to mount trials of controversial treat-
ments for low back pain [36]. Consequently, in order to
recruit adequate numbers of suitable patients, the present
study relied on advertising as well as other venues. Also,
patients with compensation claims were deliberately ex-
cluded in order to avoid potentially confounding effects of
litigation or motivation. As a result, the sample obtained
may not be representative of what might be considered the
typical patient with chronic low back pain. The patients
were less disabled and comparatively healthier. Furthermore,
dwindling interest in the present study slowed patient enroll-
ment, and despite every effort to maintain recruitment, refer-
rals became few. Accordingly, in the interests of the
participants, the study was closed and the results obtained
to date were analyzed.
Despite these difficulties, the sample obtained was appro-
priate for the nature of the study, which was an explanatory
trial: to test IDET against sham treatment. For that purpose,
the critical eligibility criterion was that the patients had
the correct indication for the procedure, which the present
sample satisfied. Furthermore, randomization succeeded in
generating two groups that were demographically and clini-
cally equivalent.
The principal findings of the study were that IDET
achieves a significantly greater improvement in pain scores
and disability than a sham treatment. In this respect, the
study answers calls for IDET to be subjected to randomized
and placebo-controlled trials.
IDET is not a universally successful treatment. Some
50% of patients do not benefit appreciably, or at all. This
feature dilutes mean scores on outcome measures and ren-
ders it difficult to demonstrate statistically significant differ-
ences without enrolling large numbers of patients.
Nevertheless, the present study found statistically significant
benefits in favor of IDET.
These features were also evident in the categorical out-
comes. Only 40% of patients treated with IDET achieved
greater than 50% relief of pain. Nevertheless, this proportion
was greater than that of patients in the sham treatment
group who achieved the same outcome. Meanwhile, few
patients treated by IDET deteriorated, but a significantly
greater proportion of sham-treated patients did so.
The present study was also compromised by what might be
called a healthy patient effect. Improvements are difficult to
demonstrate statistically if a large proportion of patients
are not particularly disabled at inception. For that reason,
the outcomes were analyzed by stratifying patients according
to baseline measures. It emerged that IDET was significantly
more effective than sham treatment in patients with poor
physical function and high disability (Table 5), but in patients
33
with low disability scores, and who were already functioning
well, the effects of IDET were indistinguishable from those
of sham treatment. This introduces the second major obser-
vation from the present study.
The sham treatment was performed in a rigorous and
disciplined manner. Perceptually, the sham treatment had all
the hallmarks of a formal, surgical treatment. It was con-
ducted in a procedure suite, with fluoroscopy operating,
and treatment devices operating. Physically, however, it
consisted of no more than the insertion of a needle into the
patients back. Yet, substantial numbers of patients benefited.
Some 38% reported an improvement in pain of greater than
20 points, and 33% reported greater than 50% improvement,
with one patient reporting complete relief of pain. Recognize
that, appropriately so, the treatment in question remained as
the only variable, while subsequent postprocedural care re-
mained as a constant for both groups.
These outcomes reveal that IDET is associated with a
substantial effect, generated by nonspecific features of the
procedure. This warns advocates of IDET to be tempered
in their enthusiasm for the procedure, but the implications
of this observation go beyond its relevance to IDET. No other
surgical procedure for discogenic low back pain has been
subjected to a placebo-controlled trial. The results of the
present study indicate that nonspecific therapeutic effects
may be a major component of the observed outcomes of
other treatments. In particular, the results of the present
study warn consumers to be wary of results claimed for new
intradiscal therapies, when those therapies have not been
subjected to placebo-controlled trials. What appear to be
reasonable or promising outcomes may be the result of no
more than the circumstances of having a procedure. The
present study warns that placebo-controlled trials of these
procedures be conducted early, rather than late, before and
lest undeserved reputations of efficacy be established, to the
detriment of the community.
In the present study, one patient in the sham treatment
group had an outstanding outcome. Her response was atypi-
cal of the other patients in that group, and she constituted
what could be defined as an outlier. In the presence of
outliers, Federal Drug Administration regulations [37] allow
that one analysis with the actual values and at least one
other analysis eliminating the outlier effect should be per-
formed and differences between their results discussed. In
the present study, with the outlier included, the results are
as reported in Tables 3 and 4. If the one outlier is censored,
not only the changes but also the mean scores at 6 months,
for VAS, Bodily Pain, Physical Functioning and Oswestry
Disability Scale, all become significantly better (p.05) in
favor of IDET. However, while drawing attention to this
possible concession, the authors have chosen not to rely on
it for drawing conclusions concerning the relative efficacy
of IDET.
The modest outcomes obtained in the present study are not
dissimilar from some of those reported by others. The
present study obtained a reduction in pain scores from 6.6
34 K.J. Pauza et al. / The Spine Journal 4 (2004) 2735
to 4.2, which is like the reduction from 6.6 to 3.7 reported
by Saal and Saal [6]. The mean change in pain of 2.4 is
barely greater than that of 1.8 reported by Derby et al. [8].
Those outcomes, however, are quite different from the im-
provement from 8.0 to 3.0, reported by Karasek and
Bogduk [11,12].
Various factors may account for differences in observed
outcomes. The samples studied may differ in psychosocial
domains. The patients may differ in the severity of disability.
Different providers may express different nonspecific effects
of treatment. Different selection criteria may be applied,
such as the requirement for postdiscography CT [2]. Opera-
tors may differ in how assiduously they perform the proce-
dure and in exactly where they place their electrode [2].
There is scope for exploring these factors in an effort to
improve the yield of IDET.
Meanwhile, the present study shows that nonspecific fac-
tors are a major determinant of the efficacy of IDET but that
its effects cannot be wholly attributed to nonspecific factors.
A needed-to-treat value of 5, for achieving 75% relief of
pain, indicates that it is a worthwhile intervention for some
highly select patients.
Acknowledgments
The authors wish to thank the physicians of The Texas
Spine and Joint Hospital, Tyler, TX, and the East Texas
Medical Center Neurological Institute, Tyler, TX, for their
unrestricted financial and institutional support of the study;
Merit Medical Systems, South Jordan, UT, for providing
digital pressure manometers for discography; Smith and
Nephew, Andover, MA, for providing SpineCath Catheters;
Ned Moore, American Professional Bracing, TX, for provid-
ing braces and Richard Silliman, MS, PT, Orthopedic
Manual Physical Therapy Institute of Dallas, Dallas, TX,
for providing postoperative rehabilitation programs.
References
[1] Saal JS, Saal JA. Management of chronic discogenic low back pain
with a thermal intradiscal catheter. A preliminary report. Spine 2000;
25:3828.
[2] Karasek M, Bogduk N. Intradiscal electrothermal annuloplasty: percu-
taneous treatment of chronic discogenic low back pain. Techniques
Reg Anesth Pain Manage 2001;5:1305.
[3] Saal JS, Saal JA. Intradiscal electrothermal annuloplasty (IDET) for
chronic disc disease: outcome assessment with minimum one year
follow-up. Proceedings of the 14th Annual Meeting of the North
American Spine Society, Chicago, October 2023, 1999. Rosemont,
IL: North American Spine Society, 1999:756.
[4] Saal JA, Saal JS. Intradiscal electrothermal annuloplasty (IDET) treat-
ment for chronic multi-level discogenic pain: prospective 1 year
follow-up outcome study. Proceedings of the 14th Annual Meeting
of the North American Spine Society, Chicago, October 2023, 1999.
Rosemont, IL: North American Spine Society, 1999:789.
[5] Derby R, ONeill C, Berquam JA, Kenderes L. Intradiscal electrother-
mal therapy by catheter: 12 month follow-up. In: Proceedings of the
7th Annual Scientific Meeting of the International Spinal Injection
Society, Las Vegas, August 1999. Daly City, CA: International Spinal
Injection Society, 1999:47.
[6] Saal JA, Saal JS. Intradiscal electrothermal treatment for chronic
discogenic low back pain. A prospective outcome study with minimum
1-year follow-up. Spine 2000;25:26227.
[7] Saal JA, Saal JS. Intradiscal electrothermal treatment for chronic
discogenic low back pain. Prospective outcome study with a minimum
2-year follow-up. Spine 2002;27:96674.
[8] Derby R, Eek B, Chen Y, ONeill C, Ryan D. Intradiscal electrothermal
annuloplasty (IDET): a novel approach for treating chronic discogenic
back pain. Neuromodulation 2000;3:828.
[9] Singh V. Intradiscal electrothermal therapy: a preliminary report.
Pain Physician 2000;3:36773.
[10] Endres SM, Fiedler GA, Larson KL. Effectiveness of intradiscal elec-
trothermal therapy in increasing function and reducing chronic low
back pain in selected patients. Wisconsin Med J 2002;101:314.
[11] Karasek M, Bogduk N. Twelve-month follow-up of a controlled trial
of intradiscal thermal anuloplasty for back pain due to internal disc
disruption. Spine 2000;25:26017.
[12] Bogduk N, Karasek M. Two-year follow-up of a controlled trial of
intradiscal electrothermal anuloplasty for chronic low back pain re-
sulting from internal disc disruption. Spine J 2002;2:34350.
[13] Anonymous. Intradiscal electrothermal therapy: conflicting results,
lively debate. The Back Letter 2001;16:133:1402.
[14] Harden N. Department editors note. Bulletin of the American Pain
Society 2001;11(4):1.
[15] Carey TS. Point of view. Spine 2002;27:9734.
[16] Karasek M, Carragee EJ, Schofferman J. Intradiscal electrothermal
therapy. SpineLine 2002;March/April:12118.
[17] Heary RF. Intradiscal electrothermal annuloplasty: the IDET proce-
dure. J Spinal Dis 2001;14:35360.
[18] The Standards of Reporting Trials Group. A proposal for structured
reporting of randomized controlled trials. JAMA 1994;272:192631.
[19] Koes BW, Bouter LM, van der Heijden GJMG. Methodological quality
of randomized clinical trials on treatment efficacy in low back pain.
Spine 1995;20:22835.
[20] Koes BW, Scholten RJPM, Mens JM, Bouter LM. Efficacy of epidural
steroid injections for low-back pain and sciatica: a systematic review
of randomized clinical trials. Pain 1995;63:27988.
[21] Koes BW, Scholten RJPM, Mens JMA, Bouter LM. Efficacy of non-
steroidal anti-inflammatory drugs for low back pain: a systematic
review of randomized clinical trials. Ann Rheum Dis 1997;56:21423.
[22] van Tulder MW, Koes BW, Bouter LM. Conservative treatment of
acute and chronic nonspecific low back pain: a systematic review
of randomized controlled trials of the most common interventions.
Spine 1997;22:212856.
[23] Merskey H, Bogduk N, editors. Classification of chronic pain. Descrip-
tions of chronic pain syndromes and definition of pain terms, 2nd ed.
Seattle: IASP Press, 1994.
[24] Endres S, Bogduk N. Practice guidelines and protocols. Lumbar disc
stimulation. Syllabus of the ISIS 9th Annual Scientific Meeting. San
Francisco: International Spinal Injection Society, 2001:145675.
[25] Digital pressure manometers. South Jordan, UT: Merit Medical
Systems.
[26] SpineCath Catheters. Andover, MA: Smith and Nephew.
[27] Huskisson EC. Measurement of pain. Lancet 1974;2:112731.
[28] Price DD, McGrath PA, Rafii A, Buckingham B. The validation of
visual analogue scales as ratio scale measures for chronic and experi-
mental pain. Pain 1983;17:4556.
[29] Carlsson AM. Assessment of chronic pain. I. Aspects of the reliability
and validity of the visual analogue scale. Pain 1983;16:87101.
[30] Chapman CR, Casey KL, Dubner R, Foley KM, Gracely RH,
Reading AE. Pain measurement: an overview. Pain 1985;22:131.
 K.J. Pauza et al. / The Spine Journal 4 (2004) 2735
[31] Ware JE, Sherbourne CD. The MOS 36-item Short-Form Health
Survey (SF-36): I. Conceptual framework and item selection. Med
Care 1992;30:47383.
[32] McHornery CA, Ware JE, Rogers WH, Raczek AE, Lu JFR. The
validity and relative precision of MOS Short-and Long-Form Health
Status Scales and Dartmouth COOP charts: results from Medical
Outcomes Study. Med Care 1992;30(suppl):MS25365.
[33] McHornery CA, Ware JE, Raczek AE. The MOS 36-item Short-Form
Health Survey (SF-36): II. Psychometric and clinical tests of validity
in measuring physical and mental health constructs. Med Care 1993;
31:24763.
Thirty-five
Years Ago
in Spine
In 1969, Kanefield et al. reported three
cases and a discussion that suggested ununited fracture as
the source of destructive lesions occasionally seen in
the spines of patients with ankylosing spondylitis [1].
Five years later, Yau and Chan reported three additional
35
[34] Fairbank JCT, Couper J, Davies JB, et al. The Oswestry low back
pain disability questionnaire. Physiotherapy 1980;72:714.
[35] Cook RJ, Sackett DL. The number needed to treat: a clinically useful
measure of treatment effect. Br Med J 1995;310:4524.
[36] Hopwood MB, Manning DEC. Lumbar epidural steroid injections:
is a clinical trial necessary or appropriate. Reg Anesth Pain Med
1999;24:57.
[37] E9 statistical principles for clinical trials: missing values and outliers,
Federal Register, vol. 63, no. 179, docket no. 97D-0174, section 5.3,
p. 49594.
cases, reviewed the existing literature and noted that the
lesions occur at stress sites at the level of previous disc
spaces and that they have the potential to heal [2]. They
concluded that these lesions were the result of stress
fractures.
References
[1] Kanefield DG, Mullins BP, Freehafer AA, et al. Destructive lesions
of the spine in rheumatoid ankylosing spondylitis. J Bone Joint
Surg (Am) 1969;51A:136975.
[2] Yau ACMC, Chan RNW. Stress fracture of the fused lumbo-dorsal
spine in ankylosing spondylitisa report of three cases. J Bone
Joint Surg (Br) 1974;56B:6817.