Rheumatol Int (2003) 23: 82–86
DOI 10.1007/s00296-002-0249-0
O R I GI N A L A R T IC L E
Pınar Borman Æ Dilek Keskin Æ Hatice Bodur
The efficacy of lumbar traction in the management
of patients with low back pain
Received: 6 May 2002 / Accepted: 15 August 2002 / Published online: 26 September 2002
Ó Springer-Verlag 2002
Abstract The literature on the efficacy of traction in
the treatment of low back pain (LBP) is conflicting.
The aim of this study was to examine its efficacy in this
disorder. Forty-two patients with at least 6 weeks of
nonspecific LBP were selected. Demographic data were
obtained. All patients completed the Oswestry disabil-
ity index (ODI) to assess disability and the 10-cm vi-
sual analog scale (VAS) for evaluation of pain. Subjects
were randomly assigned into group 1, receiving only
standard physical therapy, or group 2, receiving stan-
dard physical therapy with conventional lumbar trac-
tion. Standard physical therapy consisted of local heat,
ultrasound for the lumbar region, and an active exer-
cise program, given for ten sessions in all. The subjects
received instruction on correct posture and recom-
mended therapeutic exercises. They were reevaluated at
the end of treatment and at 3-month follow-up. The
mean outcome measures were global improvement and
satisfaction with the therapy, as well as disability by
ODI and pain by VAS. There were no group differ-
ences in terms of demographic and baseline clinical
characteristics. There was a significant reduction in
pain intensity and disability at the end of treatment in
both groups. There was complete or mild improvement
in 47.6% of group 1 and 40% of group 2. The satis-
faction rate with both treatments was more than 70%
immediately after the therapies. During the 3-month
period, the outcome measures except disability re-
mained statistically stable, with no difference amongst
groups. Disability was significantly reduced at follow-
P. Borman (&) Æ D. Keskin Æ H. Bodur
Clinic of Physical Medicine and Rehabilitation,
Numune Training and Research Hospital,
Ankara, Turkey
E-mail: pinarb@ato.org.tr
Tel.: +90-312-3103030
Fax: +90-312-3103460
P. Borman
Iran Cad. Turan Emeksiz Sok.,
5A/3 GOP, 06700 Ankara, Turkey
up in both groups. Of the patients, 51% continued with
the recommended exercises and had significantly lower
disability scores than those who did not continue with
the exercises. Pain and global improvement were also
better in this group, but the difference was not statis-
tically significant. In conclusion, no specific effect of
traction on standard physical therapy was observed in
our study group. We suggest focusing on back educa-
tion and exercise therapy in the management of
patients suffering from this chronic condition.
Keywords Exercise Æ Low back pain Æ
Physical therapy Æ Traction
Introduction
Low back pain (LBP) is one of the most common
medical problems and causes a significant amount of
disability and incapacity in different countries [1, 2].
Chronic LBP has a persistent and recurrent nature, with
major consequences for individuals and society. It rep-
resents a particularly costly sociomedical problem due to
the expenditure associated with repeated treatments,
long-term absence from work, and the need for social
support [3, 4]. Despite an increased awareness of this
condition, physicians still have difficulty managing it, as
there is no consensus about treatment.
The efficacy of many physical therapy modalities in
LBP is questionable, and the optimal treatment remains
obscured [4, 5]. One of the therapeutic options is trac-
tion, which can be combined with other physical therapy
modalities [6, 7, 8]. The literature on the efficacy of
traction as a part of physical therapy in the treatment of
LBP is conflicting. Lumbar traction has long been a
preferred method for treating lumbar disc problems, but
in light of the effectiveness of more active treatment, it is
generally not recommended in the treatment of acute
LBP [5, 7].
The aim of this study was to examine the efficacy of
traction therapy for patients with persistent LBP.

Materials and methods
Forty-two consecutive patients (28 women, 14 men) with persistent,
nonspecific LBP participated in this study. Subjects were recruited
from the outpatient service of the physical medicine and rehabili-
tation department of a large training hospital in Ankara. All
patients gave informed consent to participate.
The main inclusion criteria were: age less than 65 years, pain for
longer than 6 months, and/or pain that was recurrent. Patients with
inflammatory, infectious, malign, or metabolic disease of the spine,
pregnancy, osteoporosis, and those with spinal operations, neuro-
logical defects, and severe orthopedic, cardiovascular, or metabolic
disorders were not included in the study.
Data about demographic characteristics including age, sex,
duration of LBP, and working status were obtained. Each patient
underwent detailed physical examination at study entry and com-
pleted a self-administered Oswestry disability index (ODI) [9]
questionnaire to assess subjective disability, as well as a 10-cm vi-
sual analog scale (VAS) for evaluation of pain. The ODI is a brief
measure of the effect of LBP on daily function by explaining ten
domains with ten questions (pain, self-care, lifting, walking, sitting,
standing, sleep, sexual life, social life, and traveling) scored on an
ordinal scale [9].
All patients had received instruction on correct posture and
ergonomic principles in activities of daily living, associated with
descriptions of recommended therapeutic exercises. Pain medica-
tions were not allowed during the treatment period.
Each patient was randomly assigned to standard physical
therapy, either with or without traction. The standard physical
therapy program included hot pack, ultrasound therapy, and an
active exercise program given consecutively five times a week for a
total of ten sessions in 2 weeks. All treatments were applied on the
same day, with a few minutes’ resting time between the therapies.
Hot packs were performed for 10 min to the low back for local
superficial heat. A Sonopuls 434 (Enraf, Holland) was used (fre-
quency 1 MHz, pulse ratio 1:4, intensity 1.0 W/cm2, and pulse
duration 2 ms) for ultrasonographic therapy The ultrasound was
applied with a slow, gliding, rotatory movement over the low back
area. The patients were treated with the same apparatus for 10 min.
They had 20-min active individual exercise sessions directed by
physiotherapists which included isometric and isotonic strength-
ening exercises for the back, abdominal, and leg muscles, as well as
gentle lumbopelvic mobilization exercises.
Patients in group 1 received only a standard physical therapy
program, and group 2 received standard physical therapy in addi-
tion to lumbar traction, applied between ultrasound therapy and
exercise sessions, five times a week for ten treatments in all for
2 weeks. We preferred motorized traction, as this form can be
standardized satisfactorily.
The subjects in group 2 were treated with ten 20-min sessions,
with patients lying on the traction table in semifowler position. The
canvas braces were attached around the iliac crest and lower tho-
racic region after unlocking the sliding table top. The physiother-
apist increased the traction force to a maximum of 50% of body
weight. All patients in group 2 were treated with a similar traction
apparatus (Eltrac 439, Enraf, Holland). In all treatment sessions,
the physiotherapists were unaware of the patients’ status.
The patients were reevaluated immediately after treatment and
at 3-month follow-up. The main outcome measures of the treat-
ment were: assessment of global recovery (the ratings of recovery
were classified on a four-point scale as completely improved, mildly
improved, not changed, or not improved), assessment of satisfac-
tion (completely satisfied, somewhat satisfied, or dissatisfied on a
three-point scale), disability by ODI, and pain intensity by VAS.
Further questions about additional treatment after the therapy,
continuing with the exercise, were asked at follow-up.
The demographic and clinical characteristics of the patients
were evaluated using descriptive statistics. Pearson’s correlation
matrix was used to assess associations between variables. We
computed the difference of outcome measure between post-treat-
ment and baseline scores for each subject and compared these
83
between the two groups using Student’s t-test and the chi-squared
test. Statistical significance was set at P=0.05. All data were ana-
lyzed with the SPSS 9.0 statistical software package.
Results
There were 28 female and 14 male patients with a mean
age of 43.1±9.53 years (range 18–61). The mean pre-
treatment ODI and VAS pain scores were 28.1±10.8
(3–44) and 5.6±1.5 (2–9), respectively. There were no
differences between groups in terms of age, sex, duration
of pain, and VAS and ODI scores at entry. The demo-
graphic characteristics of the patients according to the
groups are shown in Table 1.
The majority of the patients were female (n=28).
There was statistically no significant difference between
the groups for gender. Gender seemed related to the
VAS pain scores in group 1. The VAS score of female
patients was 3.0±1.4, while that of male subjects was
1.5±1.04 (P<0.05). There was no relationship between
age and clinical variables of the patients, but older
patients tended to have nonsignificantly higher ODI
scores. There was an association between ODI and VAS
pain scores in both groups (P<0.05). No correlation
was observed between duration of disease and clinical
variables.
All patients received physical therapy either with or
without traction. For both groups, there was a signifi-
cant reduction in mean pain intensity immediately after
therapy (P<0.01), with no significant difference
amongst the groups. Changes in outcome measure-
ments, including VAS pain and ODI, are shown in Ta-
ble 2. There was also a significant reduction in self-rated
disability scores after both of the therapy groups, indi-
cating improvement on disability (P<0.01). In group 1
immediately after therapy, 47.6% of the patients were
completely or mildly improved, 28.5% were unchanged,
and 23.8% were worse, while these improvement ratios
were reported as 40%, 35%, and 25% in group 2 (Ta-
ble 3). In both groups, global patient satisfaction im-
mediately after therapy in both groups was also similar
(Table 4). Fifteen (71.4%) patients in group 1 and 17
(80%) patients in group 2 reported complete or mild
satisfaction with the therapy immediately after the
treatment period. No difference was observed in
Table 1 Demographic characteristics of the patients
Group 1 (n=21) Group 2 (n=21)
Age (years) 42.8±10.5 38.5±8.4
Sex (f:m) 13:8 15:6
Employed 8 13
Unemployed 5 4
Retired 3 1
Homemaker 8 3
Duration of 34.09±14.1 27±19.5
LBP (months)
Ratio of patients with and 13:8 14:7
without pain radiation
84

Table 2 The distribution of the scores of VAS pain and ODI be-
fore, immediately after, and 3 months after the treatment. Values in
parentheses show range
Group 1 Group 2
VAS pain before 5.6±1.7 (2–9) 5.7±1.1 (3–8)
Immediately after 3.8±1.4 (1–7) 3.8 ±1.1 (1–6)
3-month follow-up 3.6±1.7 (0–6) 4.1±1.7 (0–7)
ODI before 25.2±10.4 (3–41) 32.3±9.6 (12–44)
Immediately after 22.9±10.1 (3–43) 26.8±9.1 (4–41)
3-month follow-up 19.7±10.8 (0–32) 23.7±10.8 (6–38)
outcome measures regarding the patients with and
without radiation of pain (P>0.05). Regarding group 2,
we found no significant additional effect of traction
compared to standard physical therapy alone.
Of the participants, two patients in group 1 and one
in group 2 were lost to follow-up. The number of
dropouts at the 3-month follow-up did not differ sig-
nificantly from that of those continuing to follow-up.
After the treatment course, 65% of the patients declared
they had received no additional treatments for their
LBP. There was no significant difference between the
two groups in this respect.
During the 3-month period after treatment, the out-
come measures remained statistically stable, except for
disability. There was a significant reduction in disability
scores in both groups (P<0.05) (Table 2). Although the
patients in group 1 had better scores on ODI and VAS
pain, the difference was not significant between groups.
The VAS pain scores of group 1 decreased, while the
scores of patients in group 2 increased at follow-up, but
the difference was not statistically significant and the
mean score of VAS pain in group 2 was still lower than
at baseline. Twelve patients (63%) in group 1 and 15
(75%) in group 2 rated themselves as improved or not
changed. In groups 1 and 2, 37% and 25% of patients,
respectively, felt probably or definitely worse at follow-
up. Regarding overall satisfaction with their treatment,
57.8% of patients in group 1 and 60% in group 2 felt
completely or somewhat satisfied (Table 4). The satis-
faction decreased at 3-month follow-up in both groups,
but the difference was not statistically significant.
Overall patient satisfaction was significantly related to
pain scores but not to disability scores.
The majority of patients in both groups (nine patients
from group 1 and 11 from group 2) declared that they
had continued at least in part with the exercises ac-
cording to the instruction in the hospital. There were no
group differences. The patients in group 1 who contin-
ued with exercise had significantly lower disability scores
(P<0.05). Global improvement and VAS pain scores
were also better in patients who continued with exercise
than in those who did not, but this difference was not
statistically significant (P>0.05). The VAS pain and
ODI measures of the patients according to the status of
continuing with exercise are shown in Table 5.
Discussion
The measurement of disability is an important compo-
nent of the management of patients with LBP, as the
physical performance of patients with LBP is obviously
different from that in patients with other clinical pain
syndromes [5]. In this study, we assessed the disability
with ODI. This index is a valid and vigorous measure
and has emerged as the most commonly recommended
condition-specific outcome measure for spinal disorders
[5, 9]. In our study, most patients had higher scores
(>30, n=24), indicating higher levels of disability. Fe-
male patients tended to have higher ODI scores than
men, but the difference was not statistically significant.
Some researchers have reported consistently higher ODI
scores in women than in men, but others have not
confirmed this [5].
Chronic LBP is a complex disorder that must be
managed with a multidisciplinary approach addressing
physical and socioeconomic aspects of the illness.
Medication and physical therapy methods including
traction have proven to be useful adjuncts to an active
program of exercise and education that promotes func-
tional restoration [10]. Lumbar traction has been used
since prehistoric times for spinal disorders [11]. Its
mechanism to relieve pain seems to separate the verte-
brae, remove pressure or contact forces from injured
tissue, increase peripheral circulation by a massage ef-
fect, and reduce muscle spasm [12]. The results of pre-
vious studies examining the efficacy of lumbar traction
yielded conflicting results [4, 6, 7, 8, 13]. Few studies

Table 3 The global assessment

of global improvement in Complete or No change No improvement P*
patients of group 1 and 2, at the mild improvement (n) and worse (n)
end of the therapy and at (n)
3-month follow-up
Group 1 at the end 10 6 5 >0.05
of therapy
Group 1 at 3-month 7 5 7
follow-up
Group 2 at the end 11 6 4 >0.05
of therapy
Group 2 at 3-month 8 7 5
follow-up

*Difference between groups


Table 4 The global assessment of satisfaction in patients of group
1 and 2 at the end of the therapy and at 3-month follow-up
N patients completely N patients not
and somewhat satisfied satisfied
Group 1 at the end of 15 (71.4%) 6 (28.6%)
therapy
Group 1 at 3-month 11 (57.8%) 8 (42.1%)
follow-up
Group 2 at the end of 17 (80.9%) 4 (19%)
therapy
Group 2 at 3-month 12 (60%) 8 (40%)
follow-up
have demonstrated the relative efficacies of different
types of traction [7, 8, 14, 15, 16]. Van der Heijde et al.
[16] noted beneficial effects of traction, but their larger
trial indicated that traction is ineffective in nonspecific
LBP [7, 14]. A study comparing traction to conventional
treatments found evidence that some traction was ben-
eficial in patients with LBP and disc herniation [7]. Re-
versible effects of traction were found in some of the
medium- and long-term studies [7], but recent medical
literature and randomized controlled trials have pro-
vided evidence that conventional traction is ineffective
for patients with LBP [6, 7, 13, 14, 17, 18]. Although
there is no compelling evidence that lumbar traction is
clinically effective, no sufficient evidence of inefficacy to
discard this method was suggested, particularly in pa-
tients with lumbar discopathies [7]. The results of our
study were consistent with previous studies in that we
could observe no specific effect of traction on standard
physical therapy. The difference between the results of
our and previous studies indicating the efficacy and in-
efficacy of traction therapy may be explained by the
differences in the diagnostic categories of LBP, available
traction techniques, and methodology [13, 14, 18].
There is a lack of consensus about the approaches of
physiatrists and physical therapists on the treatment
efficacy of lumbar traction. In a recent study examining
the approaches of physical therapists, mechanical trac-
tion has been reported as an effective treatment modality
in acute and subacute LBP by 36% of the therapists [19].
In a study regarding the beliefs of physicians about
treatment efficacy for patients with LBP, physical ther-
apy was found to be the most effective treatment,
whereas roughly one fourth of the physicians believed
traction is effective [20]. Most physicians believed that
physical therapy and inpatient multidisciplinary treat-
ment programs were effective for chronic LBP [20, 21].
This can be due to the absence of clear evidence-based
clinical guidelines [17, 20, 21].
Management goals in chronic, persistent LBP are to
improve the ability to perform basic daily activities, re-
duce disability, and improve strength associated with a
reconditioning program [3, 4, 21] The treatment for
persistent nonspecific LBP varies but includes medica-
tion, physical therapy modalities, and exercise therapy
[11]. In this study, patients were not allowed medication
for their LBP. Hot pack and ultrasound therapies are
85
Table 5 The distribution of the scores of VAS pain and ODI in
patients who continued and not continued to exercises (mean±
SD). Given are the scores at follow-up
Continued with Not continued with P
exercise (n=20) exercise (n=19)
VAS pain group 1 3.8±0.4 4.1±0.3 >0.05
VAS pain group 2 4.0±2.7 4.9±1.9 >0.05
ODI group 1 23.6±11.1 28.4±14.6 <0.05
ODI group 2 22.2±11.6 26.9±9.4 <0.05
used in our patients for local superficial and deep ther-
mal therapy, respectively, in addition to exercise ses-
sions.
The physical properties of superficial and deep ther-
mal modalities differ in response to produce temperature
changes and overcome the combination of skin tolerance
and tissue thermal conductivity at different depths.
Temperature changes of a few degrees at different depths
alter nerve conduction, increase collagen extensibility,
lessen pain, and speed recovery [22]. Therefore, thermal
therapies are considered complementary to exercise
sessions and help increase the effectiveness of other
modalities.
In our study, both groups received standard physical
therapy, but patients who received only standard
physical therapy including exercise, local heat, and
ultrasound had better improvement than those who
received additional traction therapy on follow-up. No
specific effect of extra traction was observed in this
study.
The improvement of our patients may be explained
by the effect of exercise therapy. Exercise is one of the
most important rehabilitation modalities [3, 7, 8]. In our
study, patients who continued with exercises had better
improvement in disability and pain scores. Exercise may
have a significant role in clinical improvement [7].
McKenzie suggested extension exercises and Plum et al.
applied a method based on principles used in body-
building which involve intensive dynamic hyperextend-
ing back exercises [8]. These exercises were reported to
be beneficial in the treatment of patients with chronic
LBP. The improvement of our patients in both groups
may also be explained by physical therapy consisting of
local superficial heat and thermal ultrasound effects, a
proposed mechanism of alleviating pain. But most of the
previous studies indicated temporary efficacy of these
modalities [19, 20].
The mean pain score was not statistically lower at
follow-up and almost increased in our patients who
received standard physical therapy and lumbar trac-
tion, while the disability scores reduced, indicating
improvement on disability. In our study, patients with
LBP reported feeling better as a result of both therapy
programs, but they still experienced recurrence of pain
at follow-up. Perhaps other factors such as psycho-
social or environmental factors not assigned in this
study may have an effect on perceiving chronic pain
[23].
86
International studies have estimated that back pain
afflicts 65% to 80% of the population at least briefly at
some time during their lives [17, 24]. The cost of LBP has
continued to dominate the total cost of health care, with
a preponderance of this expense consumed by chronic
patients [25]. Different physical therapy programs are
often associated with vastly different implementation
costs. Therefore, chronic LBP must be treated optimally
in order to reduce this high financial and human cost
[19, 25]. Patient education and home-based exercise
seem more cost effective [25].
Our results do not provide evidence for the additional
effects of traction on traditional physical therapy in
patients with persistent, nonspecific LBP. We suggest a
focus on back education and exercise therapy in the
management of patients suffering from persistent LBP,
as these methods are efficacious and allow considerable
savings in the treatment costs of this chronic condition.
References
1. Deyo RA, Rainwille Y, Kent DL (1992) What can the history
and physical examination tell us about low back pain? JAMA
268:760–765
2. Tuzun C, Yorulmaz I, Cindas A, Vatan S (1999) Low back pain
and posture. Clin Rheumatol 18:308–312
3. Mannion AF, Müntener M, Taimela S, Dvorak J (2001)
Comparison of three active therapies for chronic low back pain:
results of a randomized clinical trial with one year follow-up.
Rheumatology 40:772–778
4. Van Tulder MW, Koes BW, Boster LM (1997) Conservative
treatment of acute and chronic nonspecific low back pain: a
systematic review of randomized controlled trials of the most
common intervention. Spine 22:2128–2156
5. Rastanen P (2001) Physical measurements and questionnaires
as diagnostic tools in chronic low back pain. J Rehabil Med
33:31–35
6. Beurskens AJ, de Vet HC, Köke AJ, Lindeman E, Regton W
et al (1995) Efficacy of traction for non-specific low back pain: a
randomized clinical trial. Lancet 346:1596–1600
7. Revel M (2000) Does traction still have a role in non-specific
low back disorders. Joint Bone Spine 67:146–149
8. Hansen FR, Bendix T, Skov P, Jensen CV, Kristanse JH et al
(1993) Intensive dynamic back muscle exercises, conventional
physiotherapy or placebo control treatment of low back pain.
Spine 18:98–107
9. Fairbank JJT, Pysent PB (2000) The Oswestry Disability Index.
Spine 25:2940–2953
10. Wheeler AH (1995) Diagnosis and management of low back
pain and sciatica. Am Fam Physician 52:1333–1341
11. Scheer SJ, Radack KL, O’Brien DR (19) Randomized con-
trolled trials in industrial low back pain relating to return to
work. Part 2: discogenic low back pain. Arch Phys Med Rehabil
77:1189–11897
12. Krause M, Refshauge KM, Dessen M, Boland R (2000) Lum-
bar spine traction, evaluation of effects and recommended ap-
plication for treatment. Man Ther 5:72–81
13. Pellechia GL (1994) Lumbar traction: a review of the literature.
J Orthop Sports Phys Ther 20:262–267
14. Beurskens AS, van der Heijden GJ, De Vet HC, Koke AJ,
Lindeman E et al (1995) The efficacy of traction for lumbar
back pain: design of a randomized clinical trial. J Manipulative
Physiol Ther 18:141–147
15. Werners R, Pynsent PB, Bulstrode CJ (1999) Randomized trial
comparing interferential therapy with motorized lumbar trac-
tion and massage in the management of low back pain in a
primary care setting. Spine 24:1579–1184
16. Van der Heijde GJ, Beurskens AJHM, Dirx MJM, Bouter LM,
Lindeman E (1995) Efficacy of lumbar traction: a randomized
clinical trial. Physiotherapy 81:29–35
17. Hart LG, Deyo RA, Cherkin DC (1995) Physician office visits
for low back pain. Frequency, clinical evaluation and treatment
patterns from a US national survey. Spine 20:11–49
18. Koes BW, Bouter LM, van der Heijden GJ (1995) Method-
ological quality of randomized clinical trials on treatment effi-
cacy in low back pain. Spine 20:228–235
19. Li LC, Bombardier C (2001) Physical therapy management of
low back pain. An explanatory survey of therapist approaches.
Physical Ther 81:1018–1027
20. Cherkin DJ, Deyo RA, Wheeler K, Ciol MA (1995) Physician
views about treating low back pain. Spine 20:1–10
21. Delitto A, Erhard RE, Bowling RW (1995) A treatment-based
classification approach to low back syndrome: identifying and
staging patients for conservative management. Phys Ther
75:470–489
22. Basford JR (1998) Physical agents. In: De Lisa JA (ed) Reha-
bilitation medicine, principles and practice. Third edn. Lippin-
cott-Raven, Philadelphia, pp 483–503
23. Nickel R, Egle UT, Eysel P, Rompe JD, Zöllner J, Hoffmann
SO (2001) Health-related quality of life and somatization in
patients with long-term low back pain. Spine 26:2271–2277
24. Frymoyer JW, Cats-Baril WL (1991) An overview of the inci-
dences and costs of low back pain. Orthop Clin North Am
22:263–271
25. Malanga GA, Nadler SF (1999) Nonoperative treatment of low
back pain. Mayo Clin Proc 74:1135–1148