The Spine Journal 8 (2008) 234242

Evidence-informed management of chronic low

back pain with traction therapy
Ralph E. Gay, MD, DC*, Jeffrey S. Brault, DO, PT
Department of Physical Medicine and Rehabilitation, Mayo Clinic, Rochester, MN, USA
Received 4 October 2007; accepted 13 October 2007

Abstract EDITORS PREFACE: The management of chronic low back pain (CLBP) has proven to be very
challenging in North America, as evidenced by its mounting socioeconomic burden. Choosing
among available nonsurgical therapies can be overwhelming for many stakeholders, including pa-
tients, health providers, policy makers, and third-party payers. Although all parties share a common
goal and wish to use limited health-care resources to support interventions most likely to result in
clinically meaningful improvements, there is often uncertainty about the most appropriate interven-
tion for a particular patient. To help understand and evaluate the various commonly used nonsurgi-
cal approaches to CLBP, the North American Spine Society has sponsored this special focus issue
of The Spine Journal, titled Evidence-Informed Management of Chronic Low Back Pain Without
Surgery. Articles in this special focus issue were contributed by leading spine practitioners and re-
searchers, who were invited to summarize the best available evidence for a particular intervention
and encouraged to make this information accessible to nonexperts. Each of the articles contains five
sections (description, theory, evidence of efficacy, harms, and summary) with common subheadings
to facilitate comparison across the 24 different interventions profiled in this special focus issue,
blending narrative and systematic review methodology as deemed appropriate by the authors. It
is hoped that articles in this special focus issue will be informative and aid in decision making
for the many stakeholders evaluating nonsurgical interventions for CLBP. 2008 Elsevier Inc.
All rights reserved.
Keywords: Chronic low back pain; Traction therapy; Chiropractics; Physical therapy

Description Terminology
Numerous nonsurgical therapies that mechanically un- Traction therapy refers to any method of separating the
load the spine have been used to treat chronic low back pain lumbar vertebrae with the primary force directed along
(CLBP) for many years. These treatments are variably the inferior-superior axis of the spine, in an attempt to treat
known as traction, distraction, or decompression therapies. CLBP.
Although new traction-based therapies are often promoted
as being superior to existing devices, their mechanical ef-
fects remain based on the principle of spinal distraction, History
thought to decompress neural structures and the interverte- Traction has been used to treat spinal disorders since at
bral disc. least 1800 BC [1]. Hippocrates (5th4th century BC) was
likely the first to devise a formal apparatus to apply spinal
FDA device/drug status: approved for these indications (VAX-D and traction [2]. By the 19th century, the traction bed was used
DRX9000). to treat scoliosis, backache, rickets, and spinal deformity,
Nothing of value received from a commercial entity related to this and traction corsets, traction chairs, and body suspension
manuscript.
* Corresponding author. Ei 2D-PM&R, 200 First Street SW, Rochester, were promoted by individual practitioners [3].
MN 55905, USA. Tel.: (507) 266-8913; fax: (507) 266-1561. Traction became a common treatment for CLBP in the
E-mail address: rgay@mayo.edu (R.E. Gay) early 20th century and opinions developed regarding how
1529-9430/08/$ see front matter 2008 Elsevier Inc. All rights reserved.
doi:10.1016/j.spinee.2007.10.025
 R.E. Gay and J.S. Brault / The Spine Journal 8 (2008) 234242 235

traction should be applied, including debates about the Practitioner, setting, and availability
ideal amount of force, degree of pull, duration of pull,
Traction therapy can be applied by chiropractors, phys-
and timing of force intervals [4]. Cyriax promoted traction
ical therapists, or medical physicians trained in the use of
for not only CLBP but also for lumbar disc lesions, theoriz-
specific traction devices for CLBP. Once the treatment pa-
ing that traction would produce negative pressure in the
rameters are established by the health provider, a clinical
disc and thereby reduce disc herniations [5]. Other investi-
assistant may also apply some of these interventions under
gators suggested that off-axis moments, such as flexion or
supervision. These interventions are widely available in the
extension, be added to axial traction to preferentially re-
United States, though specific devices may be limited to
duce back or leg pain. As a result, a group of therapies col-
lectively referred to as distraction-manipulation developed proprietary spine centers.
of which auto-traction (AT) [6] and flexion-distraction
(FD) [7] are perhaps best known. Reimbursement
Traction is reimbursed by most insurance plans under
Current Procedural Terminology code 97012, Application
Subtypes of a modality to one or more areas, traction, mechanical.
The most common classification of traction therapy is Distraction-manipulation is often billed by chiropractors
based on the duration of application, which may be 1) con- under CPT code 98940, Chiropractic manipulative treat-
tinuous (hours to days), 2) sustained (2060 minutes), or 3) ment, one or two regions. The typical cost for a session
intermittent (alternating traction and relaxation with cycles of traction therapy is $50 to $100.
of a few minutes or less) [48]. Traction therapy can also be Most insurers will cover traction therapy that is pre-
described by the direction of force, whether 1) axial, 2) po- scribed by a licensed health provider, though there may
sitional distraction, or 3) distraction-manipulation. Axial be limits on the number of sessions allowed per episode
traction limits the force to the superior-inferior (caudad- or year.
cephalad) axis of pull. Positional distraction and AT allow
the patient to determine the direction and amount of force Regulatory status
based on improvement in their symptoms. In distraction- Many traction therapy devices are regulated by FDA as
manipulation, the provider decides the direction and extent class II medical devices based on substantial equivalence to
of off-axis force (most often flexion, lateral flexion, or existing devices.
extension) to be used based on patient symptoms and toler-
ance for the treatment. The amount of recommended force
for axial traction varies but high-dose traction (30%50% Theory
of body weight) is thought to be most effective.
Mechanism of action
Several theories have been proposed to explain the pos-
General description
sible clinical benefit of traction therapy for CLBP. Distract-
Traction can be applied with the patient in most posi- ing the motion segment is thought to change the position of
tions including supine, prone, side-lying or suspended up- the nucleus pulposus relative to the posterior annulus fibro-
right or inverted. Patients receiving axial traction are sus [79] or change the disc-nerve interface [10]. These ef-
most often treated in the supine position with the knees fects are plausible based on studies examining the
and hips partly flexed. The use of a split-table (stationary kinematics of the lumbar spine during traction therapies.
upper portion and mobile lower portion) reduces the In addition to separating the vertebrae, traction has been
amount of force needed to counteract body weight and sep- shown to reduce nucleus pulposus pressure [11,12] and in-
arate the vertebrae. Typically a harness is applied to both crease foraminal area [12]. However, it is unlikely that me-
the pelvis and the chest and force is transmitted from the chanical changes observed in a prone position will be
device through the harnesses. Although originally applied sustained after a patient resumes an upright, weightbearing
by manual means or by using weights, axial traction is most posture. Any lasting clinical response to traction therapy
often applied with motorized or hydraulic systems today. would more likely be because of the effect of traction on
Patients receiving positional distraction and distraction- the mechanobiology of the motion segment or neural
manipulation are often treated in a prone or side-lying po- tissues.
sition with a special table that puts the patients spine in Complicating the issue further is that not all traction
specific postures according to the targeted tissues and de- therapies exert the same force on the spine and animal stud-
sired effect. A harness may or may not be used and often ies have found the mechanobiology of the disc to be sensi-
is applied only to the pelvis or ankles to provide axial trac- tive to the amount, frequency, and duration of loading [13].
tion whereas off-axis forces are provided by varying the It is possible that some forms of traction stimulate disc or
body posture or motion of the table. joint repair [14], whereas others promote tissue degradation
236 R.E. Gay and J.S. Brault / The Spine Journal 8 (2008) 234242

[15]. Although these variables have not been systematically Contraindications of traction therapy are largely based
examined, even in animal models, what is known regarding on the expected mechanical effects of traction on anatomi-
disc mechanobiology should alert us to the possibility that cal structures and physiologic effects on the cardiovascular
not all traction therapies are equal. If distracting the spine system (caused by using a harness or putting the patient in
can influence disc and joint mechanobiology, different an inverted position). Commonly listed contraindications
modes of traction may result in different clinical results. include spinal malignancy, spinal cord compression, local
Systematic reviews of lumbar traction therapy have typi- infection (osteomyelitis, disciitis), osteoporosis, inflamma-
cally not considered that different effects may exist based tory spondyloarthritis, acute fracture, aortic or iliac aneu-
on force and time parameters. rysm, abdominal hernia, pregnancy, severe hemorrhoids,
Traction trials have most often included patients with uncontrolled hypertension, and severe cardiovascular or
a mix of clinical presentations including back-dominant respiratory disease [8,17].
low back pain (LBP), leg-dominant LBP, or both. However, The available evidence does not define an ideal CLBP
a patient with only back-dominant LBP and no radiculop- patient for traction therapy. Because the effects of traction
athy is likely experiencing pain from a sclerotomal source, are primarily mechanical, criteria used to identifying pa-
such as facet joints or disc, whereas sciatic pain, even if tients likely to respond to other mechanical treatments such
caused by disc herniation, may be predominately of neural as spinal manipulation might be helpful. In that specific
origin. Although it is reasonable to suspect that traction case, patients with CLBP who have not responded to more
therapies may affect these conditions differently, there is conservative measures may be better candidates.
insufficient evidence to support this hypothesis.
Distraction-manipulation and positional distraction are
mechanically different than traditional traction (intermittent
or sustained). Rather than allowing forces to be dispersed Evidence of efficacy
throughout the lumbar tissues, these treatments attempt to
Review methods
concentrate them in a smaller area. AT, for example, allows
the patient to concentrate the force by finding the position The objective of this paper was to review randomized
that most relieves their pain and applying distraction in that clinical trials (RCTs) in the English literature to determine
position. Distraction-manipulation, most often used by chi- the efficacy of traction based for CLBP. Because the forces
ropractors and physical therapists, is performed on treat- and temporal patterns used with these therapies differ, we
ment tables that allow the operator to determine the also sought to characterize the literature in regard to the
moment-to-moment vector and timing of the distractive effect of different modes of traction.
force. These techniques include FD (Cox technique), Lean- A computer search of the English language Medline da-
der technique, and Saunders ActiveTrac method, among tabase from 1996 to present was completed using condition
others. specific terms developed by the Cochrane Back Review
Group for lumbar conditions. These results were combined
with the intervention search terms of traction, distrac-
Diagnostic testing required
tion, or decompression, and then with systematic
Anterior-posterior and lateral radiographs of the lumbar reviews or controlled trials. The bibliographies of
spine should be reviewed to exclude disease states such as recent systematic reviews of lumbar traction were searched
severe osteoporosis or ligamentous instability that might for additional references. The studies were organized into
compromise bone or soft-tissue integrity. If signs or symp- categories based on the mechanical parameters of the trac-
toms of neurological compromise are present, appropriate tion applied. Different modes of traction have been dis-
imaging (magnetic resonance imaging or computed tomog- cussed in previous reviews that considered whether the
raphy) should be obtained to determine the cause. If history force was sustained or intermittent, [18] and whether it
or clinical examination reveals the presence of red flags was generated by manual or motorized means [19]. In ad-
[16], appropriate diagnostic tests should be obtained. dition to intermittent or sustained force application, we
chose to consider whether the forces imparted were dis-
persed throughout the lumbar spine or if there was a logical
Indications and contraindications
effort to focus the force (Table 1). To aid the reader in re-
Lumbar traction therapies are most commonly used for membering the difference between these variables, the term
subacute LBP or CLBP with or without leg pain. There distraction has been used to refer to therapies that attempt
are no examination findings (clinical, imaging, or labora- to concentrate forces within a specific segmental level or
tory) that have been shown to differentiate patients who area, and the term traction to refer to the therapies that
are likely to benefit from traction therapies. Therefore, allow dispersion of forces throughout the lumbar spine
the decision to use traction therapy is based on the under- (although the vector of the forces might vary).
lying theories and a lack of benefit from more conservative The definition of chronic used in the studies ranged from
or less costly treatment options. 6 to 12 weeks duration. Because several studies used longer
 R.E. Gay and J.S. Brault / The Spine Journal 8 (2008) 234242 237

Table 1 (less than 25% body weight) sustained traction as treatment

Categories of traction therapies based on temporal (sustained vs for 25 patients with longer than 3 months of LBP [24]. Al-
intermittant) and force (focused vs dispersed) variables
though no differences were found, this study had a very
Sustained Intermittent small sample size and was likely underpowered. Addition-
Focused Positional distraction Distraction-manipulation ally, the operational definitions of high and low-dose trac-
Auto-traction Flexion-distraction (Cox) tion were not substantially different that one would expect
Positional distraction Leander technique
a detectable difference in clinical outcome.
Dispersed Sustained traction Intermittent traction Borman et al. randomized 42 subjects with LBP of at
Split-table traction VAX-D
least 6 weeks duration to physical therapy (hot pack, ultra-
Gravity traction DRX-9000
LTX 3000 Split-table traction sound, and active exercises) or physical therapy plus sus-
tained traction (greater than 50% body weight). After 10
treatment sessions over 2 weeks, they found no difference
between the groups in regard to visual analog scale
than 4 weeks duration of symptoms as entry criteria, we (VAS), Oswestry Disability Index score, or assessment of
chose to include these studies. Thus, the results discussed global recovery. Likewise, there were no differences at 3
herein relate to subjects with longer than 4 weeks of LBP months after treatment [25].
(subacute or chronic). Konrad et al. randomized 158 subjects with LBP with or
The computerized literature search returned 36 possible without thigh pain (13 months duration) to 4 treatment
studies and was supplemented by examining the bibliogra- groups: 1) underwater sustained traction, 2) balneotherapy
phies of 4 systematic reviews of lumbar traction [1821]. (hot mineral water baths), 3) underwater massage, or 4)
This strategy identified 24 RCTs that used traction therapy control group without treatment. All treatments were per-
in either the experimental or control arm. Almost all RCTs formed for 15 minutes three times per week for 4 weeks
included some subjects with sciatica or radiculopathy. Tri- (12 treatments). The outcome measures were analgesic
als in which leg pain (sciatica or radiculopathy) was used as use, VAS, spinal motion, and the straight-leg raise test. Af-
an inclusion criterion were excluded [9], as were non- ter the 4 weeks, the three treatment groups had improved
English papers [3]. Also excluded was one trial with subjects compared with the control group in both analgesic use
having less than 4 weeks of pain, and one with only imme- and VAS (p!.01), with no differences between interven-
diate posttreatment outcome after a single treatment. Of the tions. At 1 year, the treatment groups continued to use less
articles reviewed, 10 were found to use subjects with either analgesic medication but there were no other differences
LBP only or LBP with or without sciatic pain of at least 4 among the groups [26].
weeks duration. Tables 25 provide details of those RCTs, Another study found that auto-traction (AT) was more
separated into the four categories from Table 1. There was beneficial than sustained traction for LBP with or without
insufficient homogeneity between trials within any categor- leg pain (see positional distraction below) [27].
ical group to consider combining the data for analysis. The These RCTs suggest that sustained lumbar traction with
trials from each category of traction therapy are summa- 30% to 50% body weight is no better than low-dose (sham)
rized below. traction, mineral baths, underwater massage, or traditional
physical therapy for LBP of longer than 4 weeks duration.
Systematic reviews
Intermittent traction. Werners et al. [28] studied 152 pa-
Four systematic reviews on traction therapy were exam-
tients aged 20 to 60 years who sought treatment in a primary
ined to identify additional RCTs not returned by the elec-
care setting for their LBP. The mean duration of the current
tronic literature search; these are discussed below [1821].
episode of pain was not provided. The length of time since
LBP was first experienced was less than 5 years for only
Randomized controlled trials 35% and more than 10 years for 37%, with 15% having
Sustained traction. Beurskens et al. published two papers continuous pain since the initial onset; most patients had
on one study (short and longer term results) of sustained chronic/recurrent pain. Participants were randomized to
lumbar traction [22,23]. Subjects (n5151) with LBP of lon- six treatments over 2 to 3 weeks (mean 16 days) of either
ger than 6 weeks duration were randomized to receive 12 interferential therapy or motorized intermittent axial trac-
treatments with either high-dose (O35% body weight) or tion and motorized spinal massage. The main outcome
low-dose sustained traction (!20% of body weight). They measures were the Oswestry Disability Index and VAS.
found no differences in global perceived effect, Roland- No significant differences were found between the groups
Morris Disability Questionnaire scores or the severity of immediately after treatment or at 3 months follow-up.
the subjects main complaints at 5 weeks, 12 weeks, or 6 Sherry et al. [29] randomized 44 subjects (aged 1865
months follow-up. years) to intermittent traction with Vertebral Axial Distrac-
Van der Heijden et al. published a small pilot RCT com- tion (VAX-D) or treatment with transcutaneous electrical
paring high-dose (30%50% body weight) with low-dose nerve stimulation (TENS). Subjects were excluded if their
Table 2

238
RCTs of sustained lumbar traction for subacute LBP and chronic LBP
Participants
recruited
Reference Design Inclusion criteria Interventions (completed) Dose Follow-up Outcome measures Conclusions
[27] RCT LBP of 13 mo, Underwater N535 15 min, 3 times/wk Immediately VAS; analgesic All treatment groups used
with or without traction for 4 wk after therapy use fewer analgesics and had less
radiation to thigh Balneotherapy N538 (4 wk) and at pain compared with the control
Underwater N544 1 y (N at group (p!.01) with no
massage 1 y5170) differences between them.
No treatment N553 Analgesic used remained less in
the treatment groups at 1 y.
[24,25] RCT with LBP$6 wk Traction N577 20 min, 12 5 wk Global perceived Intention-to-treat analysis: no
blinded 35%50% body treatments in 5 wk effect differences at any time point

R.E. Gay and J.S. Brault / The Spine Journal 8 (2008) 234242
assessment Age 18 y weight Mean age 12 wk VAS for three
and above 39610 y main complaints
No prior lumbar 52%O6 mo 6 mo RMQ
traction duration
Traction!20% N574 ADL disability
body weight Age 42611 y
54%O6 mo duration
[24] RCT O3 mo LBP with Traction N513 (11) 1012 5 and 9 wk, 2 y Global perceived Only three in each group felt they
or without radiation 30%50% body treatments over effect had received sham traction.
Age 2565 y weight Age 4668 y 4 wk (3 times/wk) VAS for three Both groups improved; no
main complaints statistical differences between
82%O24 mo duration RMQ the group outcomes. (Study
Radiating ADL disability underpowered)
pain 73%
Sham N512 (11)
traction!25% Age 4768
body weight 83%O24 mo duration
Radiating pain
58%
[26] RCT $6 mo LBP with Standard N521(19) PT: 40 min Immediate Four-point global Both groups improved with
or without physical therapy 5 times/wk posttreatment recovery scale no significant differences
radiation (hot pack, ultrasound, for 2 wk (10 and at 3 mo between groups
Age!65 y active exercises) Age 42.8610.5 y sessions) ODI
Duration 34.1614.1 mo VAS
(13 with radiation)
Standard physical N521(20) PT: 40 min 5
therapy with Age 38.568.4 y times/wk for
traction $50% Duration 27619.5 mo 2 wk (10 sessions);
body weight (14 with radiation) traction 10 min
VAS5pain visual analog scale; PDI5pain disability index; RMQ5Roland-Morris Questionnaire; ODI5Oswestry Disability Index; LBP5low back pain; RCT5randomized controlled trials;
PT5Passive sustained traction; ADL5Activities of daily living
 R.E. Gay and J.S. Brault / The Spine Journal 8 (2008) 234242 239

Table 3
RCTs of intermittent lumbar traction for subacute LBP and chronic LBP
Participants
Inclusion recruited Outcome
Reference Design criteria Interventions (completed) Dose Follow-up measures Conclusions
[29] RCT Seeking Interferential N574 610 min Immediate VAS 84% follow-up;
treatment therapy treatments and 3 mo no statistically
for LBP over 3 wk posttreatment significant
Age 2060 y Mean age ODI differences
(65%O5 y 38.3 (9.4) y between groups
since Mean VAS
first LBP) 49.7(13.3)
Motorized N573 610 min
traction Mean age treatments
(10-20 kg) 39.2 (SD 9.5) y over 3 wk
Mean VAS 50.6
(SD15.1)
[30] RCT Age 1865 y, VAX-D N522 (19) 5 times/wk Immediate VAS VAX-D 68.4%
minimum Mean age for 4 wk, then and 6 mo Four-point success and
VAS 2 42 y; mean once/wk for posttreatment disability scale TENS
duration 8.4 y 4 wk (mean 21) % Success 0% success;
(range 0.25 (success 5 50% (p!.001 95%
30 y) pain CI 47.589.3%).
TENS N522 (21) 5 times/wk improvement) 10 of 13 still
Mean age for 4 wk, then success at
41 y; mean once/wk for 6 mo
duration 6.2 y 4 wk (mean 18)
(range 0.528 y)
LBP5low back pain; RCT5randomized controlled trials; VAS5pain visual analog scale; ODI5Oswestry Disability Index; VAX-D5Vertebral Axial
Decompression; TENS5Transcutaneous electrical nerve stimulation; CI5Confidence interval.

baseline VAS was less than 2. The mean duration of pain treatments). Success was defined as 50% improvement in
for all subjects was 7.3 years (range 0.2530 years). The VAS. One subject from each group was excluded from
VAX-D group had a mean duration of 8.4 years and a mean the analysis because they did not complete treatment.
baseline VAS of 5.99 (range 2.18.7). The TENS group had Two subjects in the VAX-D group were found to have been
a mean pain duration of 6.2 years and mean VAS of 5.44 entered with VAS less than 2 and were also excluded from
(range 2.78.5). All subjects were treated for 30 minutes the analysis. Thirteen of 19 VAX-D subjects had treatment
daily for 20 days then once a week for 4 weeks (24 success compared with 0 of 21 TENS subjects (68% vs.

Table 4
RCTs of positional distraction lumbar traction for subacute LBP and chronic LBP
Participants
Inclusion recruited Outcome
Reference Design criteria Interventions (completed) Dose Follow-up measures Conclusions
[28] RCT with Unremitting AT N522 for each 36 s every AT assessment VAS 5 AT subjects
crossover LBP with group. No group minute for after three crossed over
or without statistics given. 3060 min; treatments (2 improved
radiating leg Mean age 310 treatments and passive with PT);
painO1 mo; 44.6610.9 y, initially, traction 18 PT subjects
disc herniation 10 LBP only, 36 more after five crossed over
or protrusion 28 back and if improved treatments; (13 improved
on imaging PT leg/root Sustained traction nonresponders McGill pain with AT).
pain, 6 root at 35% body crossed over. Questionnaire 16 of 30 subjects
pain only. Mean weight) for 45 min; Final follow-up ODI treated with
duration 30 mo 5 treatments at 3 mo AT and 3 of
(1 mo to 20 y) initially, 5 6 subjects
additional treated with passive
if improved traction maintained
improvements
at 3 mo
RCT5randomized controlled trials; LBP5low back pain; AT5auto-traction; PT5Passive sustained traction.
240 R.E. Gay and J.S. Brault / The Spine Journal 8 (2008) 234242

Table 5
RCTs of distraction-manipulation lumbar traction for subacute LBP and chronic LBP
Participants
Inclusion recruited Outcome
Reference Design criteria Interventions (completed) Dose Follow-up measures Conclusions
[31] RCT Age 18 y and FD N5123 (110) 24 times 4 wk, 3, 6, Pain-VAS Statistically significant
above; Mean age per week and 12 mo RMQ difference in VAS favoring
LBPO3 mo 42.2 y for 4 wk FD at 5 wk (p5.01) in both
Radiculopathy 22 SF-36 intention-to-treat and completed
ATEP N5112 (87) treatment analyses; Patients
Mean age 40.9 y with moderate or severe
Radiculopathy 23 sustained pain benefited
most from FD and those
with recurrent pain
from ATEP
[33] RCT Age 18 y FD with Mean age Eight 3 wk post PDI Both groups
and above; trigger-point 51 y (14.2) treatments treatment improved; no difference
subacute LBP therapy, Mean duration over 3 wk RMQ in PDI or RMQ
(412 wk) or n554 (54) 4y
chronic LBP Sham Mean age 53 y
(O12 wk) manipulation (15.2)
and effleurage, Mean
n557 (52) duration 7 y
RCT5randomized controlled trials; LBP5low back pain; FD5flexion-distraction; ATEP5active trunk exercise program; VAS5Pain visual analog scale;
PDI5Pain disability index; RMQ5Roland-Morris Questionnaire; ODI5Oswestry Disability Index.

0%, p!.001); 10 of the 13 who improved with VAX-D (n5123) with an active trunk exercise program provided
where noted to have maintained their improved status at by physical therapists (n5112). All subjects had a primary
6 months. The authors hypothesized that the lack of im- complaint of LBP of at least 3 months duration. After 4
provement in the TENS group (improvement would be ex- weeks of treatment, both groups improved but there was
pected based on nonspecific effects alone) was possibly a statistically significant difference in pain (VAS) favoring
because of a negative placebo effect. These two studies FD (p5.01) in both intention-to-treat and completed treat-
of intermittent traction for LBP provide conflicting ment analyses. Subgroup analyses indicated that patients
evidence of benefit. with sustained, moderate to severe pain benefited most
from FD whereas those with chronic recurrent pain seemed
Positional distraction. Patients with LBP and/or leg pain to benefit most from the therapeutic exercise protocol.
commonly experience relief by assuming specific postures Overall, patient improvement was maintained at 12 months
or positions. This observation has led to the theory that spi- [31].
nal distraction in a relatively pain-free posture is preferable Hawk et al. [32] compared FD with sham manipulation
to simple axial distraction. AT is based on this theory of po- in subjects with longer than 4 weeks of LBP and found that
sitional distraction. Most AT studies have studied subjects both groups were improved 3 weeks after treatment with no
with radiculopathy. Only one RCT studied patients with difference between the groups in Pain Disability Index or
LBP with or without leg pain, although only 10 of 44 sub- RMQ.
jects had LBP only [27]. That study, a crossover RCT, com- Beyerman et al. randomized subjects to receive com-
pared AT (n522) with sustained traction with 35% body bined FD and thrust spinal manipulation or treatment with
weight (n522). Eighteen subjects who had been random- hot packs [33]. Unfortunately, the effect of FD cannot be
ized to sustained traction crossed over to AT because of isolated from spinal manipulation in that trial.
lack of improvement. Only five of the AT subjects crossed
over to sustained traction. At 3 months follow-up, 16 of 30
subjects treated with AT and 3 of 6 subjects treated with
Ongoing studies
sustained traction maintained their improvements. There
were no data presented that addressed the response of Two trials involving traction-based therapies for LBP
subjects with LBP only. were identified on clinicaltrials.gov at the time of this writ-
ing; Clinical Efficacy and Safety of Axiom Worldwide
Distraction-manipulation (FD). Although FD has been DRX9000 Axial Decompression System for Treatment of
used by chiropractors since the early 1970s, the first RCT Low Back Pain (NEMA Research, Inc.) and Chiroprac-
examining its efficacy for LBP was only recently published tic Prone Distraction for Lower Back Pain (Samueli Insti-
[30]. That study compared FD provided by chiropractors tute for Information Biology).
 R.E. Gay and J.S. Brault / The Spine Journal 8 (2008) 234242 241

Harms contractiondallowing better separation of the vertebraed

there is little evidence that they truly differ from simple
No studies have systematically enumerated the inci-
intermittent axial traction and it is unclear if muscle tension
dence rates of side effects caused by lumbar traction ther-
plays a major role in modifying vertebral displacement
apy. Increased intra-abdominal and chest pressure may
during traction.
result in transient cardiopulmonary side effects such as
Because the mechanical input of distraction-manipula-
shortness of breath or hypertension in susceptible individ-
tion therapies is different than traction, the results of RCTs
uals. Gravity traction has been associated with increased
using sustained traction or even intermittent traction should
blood pressure. There is little information in the literature
not be generalized to them unless evidence is forthcoming
regarding adverse events with traction therapy. A few trials that proves their physiological and mechanobiological ef-
have reported adverse events, the most common being ag-
fects to be the same.
gravation of the pain syndrome being treated. Of the 24 tri-
Although we systematically searched the literature, we
als reviewed by Clarke et al., 4 reported no adverse events,
did not search non-English language databases. Also, we
6 reported some events, and 14 did not mention adverse
did not independently grade the quality of these studies.
events [34]. One study reported aggravation of neurological
Most of them have been considered of poor quality in prior
signs in patients with sciatica (approximately 20% in treat-
systematic reviews. Of the studies we reviewed, only Tesio
ment and placebo groups) [35]. A case report noted sudden
and Merlo [27] and Beurskens et al. [23] were previously
progression of lumbar disc protrusion during VAX-D treat- graded as high quality [18,20,34]. At least two of the RCTs
ment [36]. No adverse events were observed in the only
we included had not been previously considered [30,32].
RCT of FD for LBP [37]. If continuous traction that re-
Properly designed RCTs are needed to determine if there
stricts patients to bed rest is used, there is potential for harm
are subgroups of LBP sufferers who benefit from specific
related to prolonged immobilization [16].
traction therapies. Patient variables that might be consid-
Serious adverse events are apparently rare judging by the
ered important include age, weight or body-mass index,
paucity of reports in the literature. Events such as cauda equi-
and specific diagnostic category (although controversial).
na syndrome or fracture are unlikely if appropriate evaluation
Treatment variables include duration and magnitude of
is performed before starting traction therapy. Nonetheless, force, direction of off-axis forces, and number of treat-
the true risk of such adverse events is unknown.
ments. Additionally, RCTs are needed to specifically eval-
Predictors of negative outcome, other than the use of
uate the proprietary forms of traction that are being
sustained traction, have not been identified in the literature.
aggressively marketed to the public. Further studies are
This may be a function of poor trial quality and the general
needed to conclusively determine whether intermittent trac-
lack of studies demonstrating efficacy.
tion, positional distraction, or distraction-manipulation is
beneficial for CLBP.

Summary
References
We systematically reviewed the English literature and
found only 10 RCTs of traction for LBP with or without [1] Kumar K. Spinal deformity and axial traction. Spine 1996;21:6535.
leg pain. As a group, this literature provides more evidence [2] Marketos SG, Skiadas P. Hippocrates. The father of spine surgery.
Spine 1999;24:13817.
against than for the use of traction therapy as a treatment [3] Shterenshis MV. The history of modern spinal traction with particular
for LBP only. reference to neural disorders. Spinal Cord 1997;35:13946.
We categorized the RCTs based on the temporal charac- [4] Hinterbuchner C. Traction. In: Basmajian JV, editor. Manipulation,
teristics of the force (sustained or intermittent) and whether traction and massage. Baltimore, MD: Williams & Wilkins, 1985:
there was an effort to concentrate the distraction force in 178.
[5] Cyriax J. Conservative treatment of lumbar disc lesions. Physiother-
specific areas of the lumbar spine. Unfortunately, there is apy 1964;50:3003.
scant data for all of the categories defined with the excep- [6] Ljunggren AE, Weber H, Larsen S. Autotraction versus manual trac-
tion of sustained traction. The preponderance of evidence tion in patients with prolapsed lumbar intervertebral discs. Scand
indicates that sustained traction is ineffective for LBP with J Rehabil Med 1984;16:11724.
[7] Cox JM, Feller J, Cox-Cid J. Distraction chiropractic adjusting: clin-
or without leg pain.
ical application and outcomes of 1,000 cases. Topics Clin Chiroprac-
The available evidence regarding intermittent traction is tic 1996;3:4559.
mixed; one RCT suggested benefit compared with TENS [8] Pellecchia GL. Lumbar traction: a review of the literature. J Orthop
and another found no benefit compared with interferential Sports Phys Ther 1994;20:2627.
therapy. Therefore, there is little evidence for or against this [9] Gudavalli MR. Biomechanics research on flexion-distraction proce-
form of traction, which is aggressively promoted in the dure. In: Cox JM, editor. Low back pain: mechanisms, diagnosis
and treatment. 6th ed.. Philadelphia, PA: Lippincott Williams & Wil-
United States health-care market (eg, VAX-D, DRX9000). kins, 1998:2638.
Although these proprietary traction machines theoretically [10] Knutsson E, Skoglund CR, Natchev E. Changes in voluntary muscle
allow the spine to be distracted without reactive muscle strength, somatosensory transmission and skin temperature
242 R.E. Gay and J.S. Brault / The Spine Journal 8 (2008) 234242

concomitant with pain relief during autotraction in patients with lum- [25] Borman P, Keskin D, Bodur H. The efficacy of lumbar traction in the
bar and sacral root lesions. Pain 1988;33:1739. management of patients with low back pain. Rheumatol Int 2003;23:
[11] Ramos G, Martin W. Effects of vertebral axial decompression on 826.
intradiscal pressure. J Neurosurg 1994;81:3503. [26] Konrad K, Tatrai T, Hunka A, Vereckei E, Korondi I. Controlled trial
[12] Gudavalli MR, Cox JM, Baker JA, Cramer G, Patwardhan AG. Inter- of balneotherapy in treatment of low back pain. Ann Rheum Dis
vertebral disc pressure changes during a chiropractic procedure. Adv 1992;51:8202.
Bioeng 1997;36:2156. [27] Tesio L, Merlo A. Autotraction versus passive traction: an open con-
[13] MacLean JJ, Lee CR, Alini M, Iatridis JC. The effects of short-term trolled study in lumbar disc herniation. Arch Phys Med Rehabil
load duration on anabolic and catabolic gene expression in the rat tail 1993;74:8716.
intervertebral disc. J Orthop Res 2005;23:11207. [28] Werners R, Pynsent PB, Bulstrode CJ. Randomized trial comparing
[14] Kroeber M, Unglaub F, Guehring T, et al. Effects of controlled dynamic interferential therapy with motorized lumbar traction and massage
disc distraction on degenerated intervertebral discs: an in vivo study on in the management of low back pain in a primary care setting. Spine
the rabbit lumbar spine model. Spine 2005;30:1817. 1999;24:157984.
[15] Iatridis JC, MacLean JJ, Ryan DA. Mechanical damage to the inter- [29] Sherry E, Kitchener P, Smart R. A prospective randomized controlled
vertebral disc annulus fibrosus subjected to tensile loading. J Bio- study of VAX-D and TENS for the treatment of chronic low back
mech 2005;38:55765. pain. Neurol Res 2001;23:7804.
[16] Bigos S, Bowyer O, Braen G. Acute low back problems in adults. [30] Gudavalli MR, Cambron JA, McGregor M, et al. A randomized clin-
Clinical practice guideline No. 14. Rockville, MD: Agency for ical trial and subgroup analysis to compare flexion-distraction with
Health Care Policy and Research. Public Health Service, US Depart- active exercise for chronic low back pain. Eur Spine J 2006;15:
ment of Health and Human Services, 1994. [AHCPR Publication No. 107082.
95-0642], 160. [31] Cambron JA, Gudavalli MR, Hedeker D, et al. One-year follow-up of
[17] Brault JS, Kappler RE, Grogg BE. Manipulation, traction and mas- a randomized clinical trial comparing flexion distraction with an ex-
sage. In: Braddom RL, editor. Physical medicine & rehabilitation. ercise program for chronic low-back pain. J Altern Complement Med
Philadelphia, PA: Elsevier, 2007:43757. 2006;12:65968.
[18] Clarke JA, van Tulder MW, Blomberg SE. Traction for low-back pain [32] Hawk C, Long CR, Rowell RM, Gudavalli MR, Jedlicka J. A ran-
with or without sciatica. Cochrane Database Syst Rev 2005;4: domized trial investigating a chiropractic manual placebo: a novel de-
CD003010. sign using standardized forces in the delivery of active and control
[19] Macario A, Pergolizzi JV. Systematic literature review of spinal de- treatments. J Altern Complement Med 2005;11:10917.
compression via motorized traction for chronic discogenic low back [33] Beyerman KL, Palmerino MB, Zohn LE, Kane GM, Foster KA. Ef-
pain. Pain Pract 2006;6:1718. ficacy of treating low back pain and dysfunction secondary to osteo-
[20] Harte AA, Baxter GD, Gracey JH. The efficacy of traction for back arthritis: chiropractic care compared with moist heat alone.
pain: a systematic review of randomized controlled trials. Arch Phys J Manipulative Physiol Ther 2006;29:10714.
Med Rehabil 2003;84:154253. [34] Clarke J, van TM, Blomberg S, de VH, van der HG, Bronfort G.
[21] van der Heijden GJ, Beurskens AJ, Koes BW, Assendelft WJ, de Traction for low back pain with or without sciatica: an updated sys-
Vet HC, Bouter LM. The efficacy of traction for back and neck pain: tematic review within the framework of the Cochrane collaboration.
a systematic, blinded review of randomized clinical trial methods. Spine 2006;31:15919.
Phys Ther 1995;75:93104. [35] Reust P, Chantraine A, Vischer TL. Treatment of lumbar sciatica with
[22] Beurskens AJ, de Vet HC, Koke AJ, et al. Efficacy of traction for or without neurological deficit using mechanical traction. A double-
non-specific low back pain: a randomised clinical trial. Lancet blind study. Schweiz Med Wochenschr 1988;118:2714.
1995;346:1596600. [36] Deen HG Jr, Rizzo TD, Fenton DS. Sudden progression of lumbar
[23] Beurskens AJ, de Vet HC, Koke AJ, et al. Efficacy of traction for disk protrusion during vertebral axial decompression traction therapy.
nonspecific low back pain. 12-week and 6-month results of a random- Mayo Clin Proc 2003;78:15546.
ized clinical trial. Spine 1997;22:275662. [37] Gudavalli MR, Cambron JA, McGregor M, et al. A randomized clin-
[24] van der Heijden GJ, Beurskens AJ, Dirx MJ, Lindeman E. Efficacy of ical trial and subgroup analysis to compare flexion-distraction with
lumbar traction: a randomised clinical trial. Physiotherapy 1995;81: active exercise for chronic low back pain. Eur Spine J 2006;15:
2935. 107082.