Tadano et al.

BMC Musculoskeletal Disorders (2019) 20:155

https://doi.org/10.1186/s12891-019-2545-9

RESEARCH ARTICLE Open Access

Lumbar mechanical traction: a

biomechanical assessment of change at the
lumbar spine
Shigeru Tadano1,7, Hideki Tanabe2, Sadao Arai3, Keiji Fujino4, Tokuhide Doi5,8 and Masami Akai6*

Abstract
Background: Lumbar traction is a traditional treatment modality for chronic low back pain (CLBP) in many
countries. However, its effectiveness has not been demonstrated in clinical practice because of the following: (1)
the lack of in vivo biomechanical confirmation of the mechanism of lumbar traction that occurs at the lumbar
spine; (2) the lack of a precise delivery system for traction force and, subsequently, the lack of reproducibility; and
(3) few randomized controlled trials proving its effectiveness and utility.
Methods: This study was planned as a preparatory experiment for a randomized clinical trial, and it aimed (1) to
examine the biomechanical change at the lumbar area under lumbar traction and confirm its reproducibility and
accuracy as a mechanical intervention, and (2) to reconfirm our clinical impression of the immediate effect of
lumbar traction. One hundred thirty-three patients with non-specific CLBP were recruited from 28 orthopaedic
clinics to undergo a biomechanical experiment and to assess and determine traction conditions for the next clinical
trial. We used two types of traction devices, which are commercially available, and incorporated other measuring
tools, such as an infrared range-finder and large extension strain gauge. The finite element method was used to
analyze the real data of pelvic girdle movement at the lumbar spine level. Self-report assessments with
representative two conditions were analyzed according to the qualitative coding method.
Results: Thirty-eight participants provided available biomechanical data. We could not measure directly what
happened in the body, but we confirmed that the distraction force lineally correlated with the movement of traction
unit at the pelvic girdle. After applying vibration force to preloading, the strain gauge showed proportional vibration of
the shifting distance without a phase lag qualitatively. FEM simulation provided at least 3.0-mm shifting distance at the
lumbar spine under 100 mm of body traction. Ninety-five participants provided a treatment diary and were classified as
no pain, improved, unchanged, and worsened. Approximately 83.2% of participants reported a positive response.
Conclusion: Lumbar traction can provide a distractive force at the lumbar spine, and patients who experience the
application of such force show an immediate response after traction.
Trial registration: University Hospital Medical Information Network - Clinical Trial Registration: UMIN-CTR000024329
(October 13, 2016).
Keywords: Chronic low back pain, Lumbar traction, Biomechanical experiment, Finite element method, Traction
conditions, Traction stiffness

* Correspondence: akai-masami@iuhw.ac.jp
6
Graduate School, International University of Health and Welfare, 4-1-26
Akasaka, Minato-ku, Tokyo 107-8402, Japan
Full list of author information is available at the end of the article

© The Author(s). 2019 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0
International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and
reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to
the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver
(http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.
Tadano et al. BMC Musculoskeletal Disorders (2019) 20:155 Page 2 of 12

Background Recently, we have been able to use a very sophisticated

Chronic low back pain (LBP) is a common and major device for lumbar traction with computerized precise
source of distress and disability among working individ- control that provides a stable mechanical condition for
uals in industrialized countries [1, 2]. In the Japanese each individual. One of the main reasons why we could
population, the prevalence of LBP is highest among not achieve definite results of lumbar traction previously
other health problems (men: 92.2 per 1000 population; could be resolved with such a device.
women: 118.2 per 1000 population) according to the sta- This study aimed (1) to examine the biomechanical
tistics by the Ministry of Health, Labour, and Welfare of change at the lumbar area under such traction and con-
Japan [3]. firm its reproducibility and accuracy as a mechanical
Among several conservative treatments for LBP, previ- intervention; and (2) to reconfirm our clinical impres-
ous studies have indicated that mechanical lumbar trac- sion of the immediate effect of lumbar traction. As a
tion is not effective for treating acute LBP [4–6]. A final research object, we plan to investigate the clinical
previous systematic review by Wegner et al. concluded effectiveness of mechanical lumbar traction on pain, dys-
that traction, either alone or in combination with other function, and quality of life in individuals with chronic
treatments, has little or no impact on pain intensity, LBP by performing a randomized controlled trial, and
functional status, global improvement, and return to the present study is a preliminary step for this.
work among people with chronic LBP [6]. This finding
might be because of (1) the lack of in vivo biomechanical Methods
confirmation of the mechanism of lumbar traction that Study design and setting
occurs at the lumbar spine; (2) the lack of a precise de- This study consisted of two parts. (1) Biomechanical
livery system for traction force and, subsequently, the experiment: We designed a biomechanical experimental
lack of reproducibility; and (3) few randomized con- system consisting of a biomechanical assessment of dis-
trolled trials proving the effectiveness and utility of lum- traction status at the lumbar area in patients with LBP,
bar traction. and finite element modeling at the lumbar spine. (2)
However, lumbar traction is still a widely accepted, Clinical qualitative study: We administered a
popular treatment modality for patients with chronic self-report assessment to participants in order to recon-
LBP [7, 8]. Many clinicians and physical therapists con- firm the immediate effect of lumbar traction, which is
tinue to use it for chronic nonspecific LBP too. Indeed, based on our clinical impression, as well as to reconfirm
expert clinical opinions, theoretical models, and research no adverse event.
suggest that certain patients with LBP respond positively The current study was part of a comprehensive clinical
to lumbar traction [9–11]. Since many patients with LBP trial that was registered in the University Hospital Medical
are not indicated for surgical intervention, conservative Information Network-Clinical Trial Registration
management for patients with LBP is important in clin- (UMIN-CTR 000024329, date opened: October 13, 2016).
ical practice. Further evidence is needed to support the
effectiveness of lumbar traction. Participants
Supposed beneficial effects of lumbar traction are as We recruited participants from six outpatient clinics that
follows: used two types of traction devices (MINATO Medical
Science, ST-2 L/2CL and OG Wellness Technologies,
 extension of the soft part tissue around the facet OL-6500/6000).
joint,
 displaced correction of the intervertebral discs and Inclusion and exclusion criteria
facet joint, The inclusion and exclusion criteria were determined ac-
 separation of the facet joint, cording to a previous clinical trial on chronic LBP [12].
 expansion in the intervertebral foramen, Participants aged 20–65 years who consulted with their
 decreased pressure of the intervertebral discs, orthopedic surgeons because of a complaint of
 extension of the anterior and posterior longitudinal non-specific chronic LBP for more than 3 months’ dur-
ligaments around the vertebral body, ation; those who were able to give informed written con-
 reduction of prolapsed discs, sent; those without neurologic deficits; and those who
 relaxation of muscle spasm by stretching, satisfied the following criteria: angle of more than 70°
 improvement of blood circulation, and during the straight-leg-raising test, negative femoral
 psychological effect. nerve stretching test result, no superficial sensory defi-
cits, and no muscle weakness less than 4/5 according to
However, some of these in vivo changes are difficult to the muscle manual testing were enrolled in this study.
prove directly. LBP was defined as pain localized below the L1 spinal
Tadano et al. BMC Musculoskeletal Disorders (2019) 20:155 Page 3 of 12

process and above the inferior gluteal folds without sci- Clinical Orthopaedic Association approved the protocol
atica (radicular pain). Chronological fluctuation of pain of the present clinical experiment (approval number:
was not considered if the pain itself continued for more 2015–01).
than 3 months.
Exclusion criteria were patients who had LBP due to Description of the experiment
tumors, infections, or fractures; previous back surgery; Device for delivering precise traction force
severe osteoporosis; psychiatric disorders, such as de- We used two types of traction devices that are commer-
pression or others; liver and renal dysfunction; preg- cially available as the same category of classification
nancy; medication for cardiac failure; a history of (MINATO Medical Science, ST-2 L/2CL and OG Well-
cerebrovascular accident and/or myocardial infarction ness Technologies, OL-6500/6000). Both device consist
within 6 months before the day of agreement to enter of two main parts: the holding part that holds the upper
the trial; and were not suitable for traction according to body and the moving part that maintains a uniform
the attending physician. 90°-90° position of the lower extremities (Fig. 1). The
upper body unit automatically measures the height of
Ethical statements the arm pit using an adjustable holding arm to keep the
Patients’ attending physicians provided an information participant in the sitting position. The lower body unit
leaflet about the experiment, completed the patient in- secures the participant in 90° of hip flexion with a pelvic
formation sheets in order to determine eligibility based girdle belt, and the thigh lengths with 90° flexion of the
on the inclusion and exclusion criteria, and obtained knee joints is maintained by lifting the shin component.
written informed consent to participate from each pa- These two main components are moved separately on
tient. The institutional review board of the Japanese the rail via an actuator connected to a load cell.

Fig. 1 Drawing schema of the traction device used in this study (the bottom of Fig. 1 is adopted from the pamphlet for the ST-2 L/2CL with
permission from MINATO Medical Science). The device consists of two main parts: the holding part that holds the upper body and the moving
part that maintains a uniform 90°-90° position of the lower extremities. Serial clinical pictures of the application of lumbar traction using the
device (ST-2 L/2CL)
Tadano et al. BMC Musculoskeletal Disorders (2019) 20:155 Page 4 of 12

After the participant laid down on the device, it took (2) The main measuring systems monitored the
73 s to start the active traction mode. During this phase, movement of the traction unit holding the pelvic
the load cell voltage indicated the same pattern of girdle belt from two aspects: the loading pattern
change in each case. (measured by the machine itself ) and distraction
First, we conducted the biomechanical experiment using distance (measured by the infrared range-finder).
the ST-2 L/2CL by incorporating other measuring tools The device delivered a regulated traction force be-
such as an infrared range-finder (SHARP, GP2D12), a tween the two main parts via a feedback mechanism
strain gauge for large deformation (KYOWA, against stress-relaxation due to viscoelastic proper-
KFEM-10-120), surface electromyogram (HARADA Elec- ties of body tissues.
tronic, 8-ch EMG Telemeter), electrocardiogram (HAR- (3) Upper body holding at the armpit by the adjustable
ADA Electronic, 8-ch EMG Telemeter), and analog/digital arm worked to counter the lumbar traction.
convertor (KYOWA, PCD320A, PCD300A). Figure 2 Technicians who had lecture with the staff of Dr.
shows the block diagram of the measuring system. Sec- Tadano’s laboratory attended each clinic to add the
ond, we performed the clinical experiment using the ST-2 special devices to the original traction unit. They
L/2CL and OL-6500/6000 in order to calculate in vivo also supervised the necessary procedures that staff
mechanical stiffness and find the difference in traction of clinics used to apply the strain gauge system.
modes. After exposing the participant’s back, the staff
The features of the device were as follows. cleaned up the skin with alcohol, pasted the strain
gauge with glue, and connected the electrical lines.
(1) The participant’s posture during traction was These procedures were performed in the annex
maintained in hip flexion of 90° with minimal room next to where the traction devices were
lumbar lordosis of the moving lower body part. The located.
length of the thigh varied in each case and was The participants in the biomechanical experiment
adjusted by 90° of knee flexion. were subjected to the following.

Fig. 2 Block diagram of the current measurement system

Tadano et al. BMC Musculoskeletal Disorders (2019) 20:155 Page 5 of 12

(4) A strain gauge was attached at the surface of their traction used in the present clinical setting (Table 1). We
lower back with the help of adhesive glue, and it evaluated various experimental conditions as a prelimin-
was connected to a static distortion meter (Fig. 3). ary study. When we shifted from one towing condition to
(5) A surface electromyogram at the three levels of another one, we allowed at least 20 min of rest as a wash-
both paravertebral muscles and an out time. In the clinical assessment, at least two days were
electrocardiogram were used to confirm interposed between the different traction conditions.
participants’ relaxation status. For the loads of traction, 40% of the body weight was
Data generated from these instruments were used as the preload. Ten minutes was used as the trac-
automatically processed through an analog-to- tion treatment time.
digital convertor and recorded by a computer as Towing A1 = Vibration function at a frequency of 0.1
synchronized data. Hz was added to the preload (traction for 10 min). The
vibration amplitude load was 30% of the preload (ampli-
We measured the following parameters with this sys- tude between 100 and 70%). This traction was repeated
tem: (1) the traction unit loading pattern, (2) traction for 60 s and suspension for 10 s.
unit shifting distance; (3) in vivo extension distance at Towing A = The vibration amplitude load was 30% of
the skin surface over the lumbar spine, (4) heart beat the preload. This traction was repeated for 30 s and sus-
rate, and (5) surface electromyographic activities. pension for 5 s.
Towing A2 = The vibration amplitude load was 20% of
Traction loading pattern the preload. This traction was repeated for 30 s and sus-
We set two types of loading patterns as a cross-over de- pension for 5 s.
sign, which were applied randomly, with a towing condi- Towing B = Traction was applied for 30 s and suspen-
tion determined according to the most popular method of sion for 5 s.
After setting the various measuring systems, we ap-
plied one set of trial traction as the adjustment, and then
performed towing conditions A and B three times each
for two weeks.

Maintaining strain gauge and the measuring conditions

Initially, it was difficult to maintain adhesion of the strain
gauge to the skin of the lumbar spine. During distractive
application of up to 10 min, some cases failed to maintain
the experimental conditions. However, this issue could be
controlled for by using tape to reinforce placement of the
gauge. Additionally, the electrode wiring clip protruded
outside, and there was a possibility that it could come off
during towing. During the experiments, the participants
wore t-shirts so that the electrode parts did not touch the
backrest of the towing machine directly. Occasionally, the
electrode clip became pulled off during towing and was
unable to obtain a measurement, but in most cases, the
myoelectric potential was measured without problems.
The traction delivery chart on display showed some
perturbation when the participants strained themselves.
Electromyographic activities could be a good indicator
for relaxation of the body muscles.
The criteria of successive data recording were dependent
on the responsiveness of the infrared range-finder, in which
the curve shifted to the + side during the traction phase
and to the – side during the relaxation phase.

Fig. 3 Photograph of a participant’s back. The photograph shows Finite element model simulation
the surface electrodes (Ch. 1–4) for electromyography attached to Computer simulation of deformation at the lumbar spine
the paravertebral muscles and the strain gauge attached to the
level under lumbar traction force was performed using
lumbar spine
FEM software (ANSYS 15.0 ANSYS. Inc.). Musculoskeletal
Tadano et al. BMC Musculoskeletal Disorders (2019) 20:155 Page 6 of 12

Table 1 Baseline characteristics of participants of biomechanical experiment and clinical study

a b b
Devices N Gender Age Age Towing mode
Male: Female (Mean) (SD) traction for 10 min
Biomechanical ST-2 L/2CL 14 9:5 35.1 11.1 A1, B
10 4:6 47.7 9.7 A, B
OL-6500/6000 7 1:6 44.3 11.5 A, B
7 3:4 40.7 10.4 A, A2
net 38 17:21 41.1 11.7
Clinical A→B 49 27:22 42.2 12.6 A, B
B→A 46 19:27 41.5 13.6 A, B
a
MINATO Medical Science, ST-2 L/2CL and OG Wellness Technologies, OL-6500/6000
Towing A = Vibration amplitude load was 30% of the preload. It’s repeated for traction 30 s and suspension 5 s
Towing A1 = Vibration function at the frequency of 0.1 Hz was added onto the preload. Vibration amplitude load was 30% of the preload. It’s repeated for traction
60 s and suspension 10 s
Towing A2 = Vibration amplitude load was 20% of the preload. It’s repeated for traction 30 s and suspension 5 s
Towing B = Traction 30 s and suspension 5 s
b
According to t-test for equality of means between A → B and B → A, age is not significant with mean difference 0.768 (95% Confidence Interval of the Difference:
− 4.555 to 6.091), p = 0.775
Also, in Chi-Square, there was no significant difference of the male: female ratio between A → B and B → A, p = 0.179

modeling of the whole body sitting in the traction device Statistical analysis
was performed using OpenSim 2.4 (National Center for Biomechanical parameters (mean values) were analyzed
Simulation in Rehabilitation Research, Stanford University) with the paired t-test to determine differences between
and CAD software (Geomagic for Solidwork). A whole towing modes A group and B. The 95% confidence inter-
body skeletal model positioned in the traction device was val (CI) of the population mean of the shifting distance,
distorted in the same manner as in clinical treatment. and correlation coefficient between traction stiffness
(TS) and age were calculated.
Treatment diary As for the clinical analysis of the treatment diary, the
The participants maintained a treatment diary about their chi-square test was used to determine differences
conditions after each intervention for two weeks. This of female-to-male ratio between A → B and B → A.
diary was requested to confirm that the device operated Other data such as age, height, weight, and body mass
without any problems and to collect self-report assess- index were analyzed using the t-test for equality of
ments by participants. They provided brief comments means between A → B and B → A. The Mantel-Haenszel
about their health status each day in their treatment diary. chi-square for 2 x r tables was used to evaluate the re-
Since we added a few parts to the commercial products sults of the self-report assessments for each traction
and modified the traction mode, we mainly investigated mode. All statistical analyses were performed with SPSS
the patients’ perception on the newly introduced condi- Statistics 17.0 for Windows (SPSS Inc.).
tion such as the vibration traction method and whether
they experienced any inconveniences. Statement of the location where the work was performed
The participants’ perspective on health status before The biomechanical research including FEM simulation
and after lumbar traction, which focused on 6 days of was performed in the laboratory of Bio-Mechanical
the actual traction trial, was analyzed using a summa- Design, Division of Human Mechanical Systems and
rized tabulation of their experiences. We used the mixed Design, Faculty of Engineering, Hokkaido University,
method with qualitative categorization and quantitative and six orthopedic clinics.
calculation. The clinical study was conducted in 28 orthopedic
An independent researcher read the participants’ com- clinics (see acknowledgements).
ments for each day and classified each one into four cat-
egories: no pain, improved, unchanged, and worsened. Results
Coding was used to rate the content of the diary over Available data acquisition
three consecutive days in order to determine a represen- For the biomechanical experiment, 38 adults aged 20–
tative health status: “n” was assigned to no pain, “i” to 59 years from six orthopedic clinics participated. The
improved, “u” to unchanged, and “w” to worsened. For ST-2 L/2CL was used in 24 cases and OL-6500/6000 in
example, a figure sequence “uui” meant unchanged – 14 cases that provided biomechanical data for calculat-
unchanged – improved for three days per week. ing stiffness and others. The remaining participants
Tadano et al. BMC Musculoskeletal Disorders (2019) 20:155 Page 7 of 12

submitted their clinical assessment of the effectiveness days. Typical comments in the diary were as follows: “I
of lumbar traction to determine the optimal traction feel so much better,” “The pain was relieved,” “My back
condition (Table 1). got lighter,” “It got easier to move,” and “The numbness
Subsequently, we collected biomechanical data from became light.” The results are shown in Table 2.
14 cases for A1 mode, 24 for A, 31 for B, and 7 for A2.
The strain data were also available for 14 cases for A1 Results of each parameter
mode, 8 for A, 22 for B, and 4 for A2 (Figs. 4 and 5). The 95% CIs of the population mean of the distance of the
For the clinical experiment, A to B traction assignment traction unit were 110.9–134.0 mm in A1 mode, 110.1–
consisted of 49 cases, whereas B to A consisted of 46 124.6 mm in B, 103.9–123.5 mm in A, and 98.5–148.8 mm
cases. We collected 570 health status assessment codes, in A2. The 95% CIs of the strain gauge distance were 1.6–
each after traction condition and converted them into 190 4.3 mm in A1 mode, 2.6–6.7 mm in A, and 1.9–3.7 mm in B.
cases with coding of diary content for three consecutive The strain gauge was able to determine the linear re-
sponse to vibration movement at the skin surface of the
lumbar spine during the first step. However, it was diffi-
cult to quantify the data of the strain gauge, which may
be due to the variations in each subject as well as the
viscoelasticity of tissue around the armpit or pelvic gir-
dle at the body holding and other sites that masked the
actual fluctuation of living tissue.

Traction stiffness
TS (N/mm) were calculated using the first peak value of
distractive load and shifting distance. In towing A1 with
the ST-2 L/2CL, the TS value was 1.96 ± 0.34 N/mm
(average value [AV] ± standard deviation [SD]), and it
was 2.09 ± 0.11 N/mm in towing A with the OL-6500/
6000. In towing B, the TS values were 2.00 ± 0.35 N/mm
and 1.94 ± 0.29 N/mm with the ST-2 L/2CL and
OL-6500/6000, respectively. There was no statistically
significant difference between the two devices (p= 0.430
and p= 0.656, t-test). There was no relationship between
TS and age, as indicated in towing conditions A1 (r = −
0.175), A (r = − 0.300), and B (r = − 0.075).
In the intra-participant comparison, however, the TS
value in towing A1 was 1.93 ± 0.09 N/mm (AV ± SD), and it
was 1.83 ± 0.08 N/mm in towing B (p= 0.001, paired t-test).
The TS value in towing A was 2.09 ± 0.11 N/mm, and it
was 1.87 ± 0.05 N/mm in towing B (p= 0.080, paired t-test).
There was no difference between towing conditions A and
A2 (p = 0.231, paired t-test). In each participant, the differ-
ence of traction mode produced distinct responses.

Finite element method simulation

In FEM stimulation, when the kyphosis of the entire spine
is fixed to the shoulder and the pelvis are under the same
condition as the present device, and towing distance is 100
mm, the lumbar region (L1–5) becomes targeted part
under distraction. To confirm whether displacement of the
spine was driven to a degree, we used the standard whole
skeleton shape model (Open Sim Model), which has a
Fig. 4 Graphic results of biomechanical measurement. Although physiological curvature (anterior lordosis), as seen in the
there is a proportional relationship between load and distance (i, e., pre-tow drawing of Fig. 6. In addition, the bone part of the
traction stiffness), the relationship between load and strain gauge or
spine had a rigid body that does not become deformed, and
between distance and strain gauge is unclear
the disc had an isotropic elastic body (elastic modulus 500
Tadano et al. BMC Musculoskeletal Disorders (2019) 20:155 Page 8 of 12

Fig. 5 An example of two different towing patterns. The strain gauge was able to determine the linear response to vibration movement at the
skin surface of the lumbar spine

MPa, Poisson ratio 0.3). We ignored any tissue or interver- body traction between the holding arm and pelvic girdle
tebral joint actions such as movement of the ligaments sur- belt. The maximum strain value was 0.55 at the L1/L2
rounding the spine. All deformation of the spine due to intervertebral disc.
towing occurs in the intervertebral disc. Under these condi-
tions, this analysis confirmed the local deformation (inter-
vertebral disc) against the deformation of the entire spinal Other findings
column, and the result did not affect the magnitude of the Voluntary muscle contraction or movement of the trunk
elastic modulus of the intervertebral disc. by the participants resulted in turbulence in the traction
According to FEM simulation, 3-mm displacement of pattern. Therefore, participants were asked to relax dur-
the lumbar spine region was obtained under 100 mm of ing the experimental period.
Tadano et al. BMC Musculoskeletal Disorders (2019) 20:155 Page 9 of 12

Table 2 Self-report assessment of lumbar traction according to The final judgment of each three-day combination was
qualitative analysis improved with nnn, nni, nnu, nin, nii, inn, ini, iin, iii, iiu,
Assessment for 3 days iui, iuu, iwi, unn, unu, uin, uii, uiu, uun, and uui; un-
Improved Changeless Worsened total changed with iwu, uuu, wnu, and wuu; and worse with
Traction A 78 11 6 95 iuw, uiw, uuw, uww, wwu, and www. Both A mode and
Traction B 79 12 4 95
B mode showed improvement in 78 (82.1%) and 79
(83.2%) cases (Table 2). There was no difference between
total 157 23 10 190
the two modes (p = 0.682). We confirmed immediate
Mantel-Haenszel Chi-Square for 2 x r table. QS (Mean score
statistic) = 0.168, p = 0.682
utility or change by lumbar traction, which was compat-
Both A mode and B mode showed improvement in 78 (82.1%) and 79 (83.2%). ible with the clinicians’ impression in daily practice.
There was no difference between two modes (p = 0.682) Additionally, there were no cases of stopping and no
adverse events.
Self-report treatment diary
The coding of diary content emerged in 30 of 64 combi- Discussion
nations (=4 × 4 × 4). The frequencies of occurrence of Despite the long history of use of lumbar traction in clin-
these combination in 190 cases were as follows: nnn, 11; ical practice, the evidence of its effectiveness has not been
nni, 1; nnu, 3; nin, 4; nii, 4; inn, 7; ini, 1; iin, 10; iii, 30; documented and proved [13]. The Cochrane Database of
iiu, 3; iui, 5; iuu, 14; iuw, 1; iwi, 1; iwu, 2; unn, 3; unu, 2; Systemic Review has indicated that traction may make lit-
uin, 2; uii, 25; uiu, 9; uiw, 1; uun, 6; uui, 16; uuu, 19; tle or no difference in pain intensity, functional status, and
uuw, 2; uww, 1; wnu, 1; wuu, 1; wwu, 3; and www, 2. global improvement or return to work when compared to

Fig. 6 Finite element modeling. Under finite element method simulation, 3-mm displacement of the lumbar spine region is obtained under 100
mm of body traction between the holding arm and pelvic girdle belt. The maximum strain value is 0.55 at the L1/L2 intervertebral disc
Tadano et al. BMC Musculoskeletal Disorders (2019) 20:155 Page 10 of 12

placebo, sham traction, or no treatment [4, 6]. To date, subgroup of patients rather than identifying the mechan-
the use of traction as treatment for non-specific LBP is ism of action. After establishing clinical utility, we think it
not supported by the best available evidence. is possible to quantitatively examine the mechanism of ac-
tion of lumbar traction. According to the system control
Interpretation of the current experiment theory, transfer function provides specific characteristics
of the targeted system, which is the lumbar spine in this
(1) The concept of traction device used in the current case. Between input as a loading force and output as a dis-
study is a link between the upper and lower body tractive distance, we calculated TS. Even after demonstrat-
segments, which are connected by the lumbar ing the reproducibility of the traction procedures, we still
spine. The upper body unit holds the body with face the lack of a “good” control (placebo, sham, or null)
holding arms, and the lower body one, which for lumbar traction that is clinically relevant.
moves on the rail, provides the constant distraction
force regulated by the computer against stress Bridging the gap to clinical effectiveness
relaxation of human tissue. Subsequently, the Several previous studies have attempted to assess the
device is uniformly able to provide machine- biomechanical changes to the lumbar spine, particularly
delivered traction to the lumbar spine with minimal the height of discs, following the lumbar traction pro-
lordosis. The reasons for the controversial results of cedure [15–17]. Furthermore, researchers have eagerly
the effectiveness of lumbar traction are mainly de- sought a precise measuring method to detect the
rived from the lack of availability to maintain con- changes, including in degenerative models [18–20].
stant interventional procedures with a conventional However, even after confirming mechanical loading,
traction device. These potential factors include posi- there may be variations in the responsiveness of patient
tioning of the participant, application method of the groups. Previously, researchers have shown such varia-
distraction force, and, sequentially, the lack of con- tions in responsiveness to lumbar traction [10, 11, 21].
firmation of mechanical loading. With newly devel- By combining the results of a clinical study and FEM
oped traction devices, we can more readily measure simulation, we could confirm some mechanical actions at
variables or control these factors. the lumbar spine in the current study. We have already
(2) In this study, we qualitatively documented the prepared some measurement scales for clinical evaluations
distraction at the lumbar spine with reproducibility, that have already undergone psychometric standardization
although the accuracy of the strain gauge on the [4, 12, 22, 23]. If we include the clinical effectiveness of
lumbar skin was not quantitatively sufficient to lumbar traction in a well-designed clinical trial, it is pos-
identify the fine movement of the pelvic traction sible to have sufficient evidence of the effectiveness of
device. Although distance measurement on the lumbar traction based on the current biomechanical
body surface has been considered difficult until now study. Actual loading patterns or the manner in which the
[14], the qualitative aspect on a drawing chart was lumbar spine experienced loading should be investigated
reasonably supportive for our research. after determining clinical effectiveness.
(3) The current experiment was the first stage of our In this study, we added vibrating force to the preload
project to examine the effect of lumbar traction. In (corresponding to 40% of the body weight) because some
the next stage, a randomized controlled trial of prior articles have reported the usefulness of vibration
lumbar traction based on patient-centered outcome on LBP, and we think that increasing sensory input
and the results of available traction conditions will could have a positive effect [24, 25].
be needed. The effectiveness of lumbar traction
should be evaluated in a combination of a biomech- Study limitations
anical study, FEM simulation, and clinical trial. Identifying an appropriate loading mode may still be an
(4) We could obtain a rough figure of shifting distance essential step for ascertaining the clinical utility of lum-
at the lumbar spine using FEM simulation. bar traction. The measurement of distraction distance it-
However, the validity of the calculated figure should self simply relies on a computer simulation conducted
be checked in a future experiment. by FEM. We only assessed the distance on the lumbar
skin and did not directly assess the shift of discs or ver-
Our current research direction tebral bodies. However, the distraction load is applied
In this study, we wanted to clarify whether lumbar trac- through the upper and lower body parts that have
tion is clinically effective in reducing LBP and determine moved on the horizontal rail. As the position of the lum-
how lumbar traction affects the body of patients with LBP bar spine maintains minimal lordosis in each case, shift-
(mechanism). Given the hesitation to use traction, we first ing distance on the skin parallelly reflects positions of
sought to determine the clinical effectiveness even in a bony elements in deeper layers.
Tadano et al. BMC Musculoskeletal Disorders (2019) 20:155 Page 11 of 12

Conclusions Ethics approval and consent to participate

The current study, which combined a biomechanical ex- All procedures performed in studies involving human participants were
conducted in accordance with the ethical standards of the institutional and/
periment with FEM simulation and analysis of patients’ or national research regulations, and with the 1964 Helsinki declaration and
perspective, found that lumbar traction operates as an its later amendments or comparable ethical standards. Written informed
actual mechanical intervention therapy for patients with consent was obtained from all participants, and the study protocol was
approved by the institutional review board of the Japanese Clinical
chronic LBP, and it provided the possibility of an imme- Orthopaedic Association (approval number: 2015–01).
diate effect after traction.
Consent for publication
Abbreviations Not applicable.
A/D convertor: analog-to-digital converter; AV: average value;
CAD: computer-aided design; CI: confidence interval; FEM: finite element
Competing interests
model; LBP: low back pain; SD: standard deviation; TS: traction stiffness
The funding body or device providers had no role in the study design, data
collection, data analysis, data interpretation, or writing of the manuscript.
Acknowledgements
We thank the orthopedic surgeons, their staff, and the patients who willingly
participated in this study. The following clinics that used traction devices Publisher’s Note
participated in this experiment: Springer Nature remains neutral with regard to jurisdictional claims in
published maps and institutional affiliations.
Arai Orthopaedic Clinic (MINATO),
Fujino Orthopaedic Clinic (MINATO), Author details
Tanabe Orthopaedic Clinic (MINATO), 1
Division of Human Mechanical Systems and Design, Faculty of Engineering,
Miyake Orthopaedic Clinic (MINATO), Hokkaido University, Kita-13, Nishi -8, Kita-ku, Sapporo-shi, Hokkaido
Harada Orthopaedic Hospital (OG), and. 060-8628, Japan. 2Tanabe Orthopaedic Clinic, 3-2-16 Narimasu, Itabashi-ku,
Osada Orthopaedic Clinic (OG). Tokyo 175-0094, Japan. 3Arai Orthopaedic Clinic, 1-19-7 Asumigaoka
Midori-ku, Chiba-shi, Chiba 267-0066, Japan. 4Fujino Orthopaedic Clinic,
We also greatly appreciate the following 28 hospitals and clinics that 2-15-12 Johoku, Naka-ku, Hamamatsu-shi, Shizuoka 432-8011, Japan.
5
participated in this project: Geriatric Care Facility Excellent Care Shizu, 1316-1 Kami-Shizu, Sakura-shi,
Akashi Orthopedic Clinic, Arai Orthopaedics Clinic, Chosi Clinic, Fujino Orthopedic Chiba 285-0846, Japan. 6Graduate School, International University of Health
Clinic, Fujino Orthopaedics, Harada Orthopedic Hospital, Hasegawa Orthopaedic & and Welfare, 4-1-26 Akasaka, Minato-ku, Tokyo 107-8402, Japan. 7Present
Rehabilitation, Iizuka Orthopaedic Clinic, Imazato Inagaki Orthopaedics, Karaki address: National Institute of Technology, Hakodate College, 14-1,
Orthopaedics, Kazusa Orthopaedics, Koike Orthopedic Clinic, Kubota Clinic, Tokura-cho, Hakodate-shi, Hokkaido 042-8501, Japan. 8Present address:
Matsuda Orthopedics Clinic, Minezaki Orthopaedic Clinic, Miyake Orthopaedic Geriatric Care Facilities Tomisato Tokushu-en, 1-1-1 Hiyoshi-dai, Tomisato-shi,
Clinic, Miyata Clinic, Nakada Hospital, Narashinodai Orthopedics & Internal Chiba 286-0201, Japan.
Medicine Clinic, Noguchi Hospital, Osada Orthopedic Clinic, Ryusoh Orthopaedic
Hospital, Suzuki Orthopaedic Clinic, Tana Orthopedic Clinic, Tanabe Orthopaedic Received: 21 August 2018 Accepted: 28 March 2019
Clinic, Uehara Orthopedic Clinic, Ueno Orthopaedics, and Wada Clinic.
Additionally, we thank the following researchers of the laboratory of
Biomechanical Design, Faculty of Engineering, Hokkaido University. They References
helped with part of the finite element model simulation, data analysis, and 1. Freburger JK, Holmes GM, Agans RP, Jackman AM, Darter JD, Wallace AS,
pilot experiments in our study: Castel LD, Kalsbeek WD, Carey TS. The rising prevalence of chronic low back
pain. Arch Intern Med. 2009;169(3):251–8.
Satoshi YAMADA, Dr. Eng., 2. Ono R, Yamazaki S, Takegami M, Suzukamo Y, Konno S, Kikuchi S, Fukuhara S.
Keeratihattayakorn SARAN, Dr. Eng., Patient-reported disability in the general Japanese population was associated
Hayato SUZUKI, Dr. Eng., and. with medical care visits for low back pain, regardless of pain intensity. J Orthop
Yumi YAMAZAKI. Sci. 2015;20(4):742–9.
3. Graphical Review of Japanese Household, from Comprehensive Survey of
Funding Living conditions. 2013. Available at: http://www.mhlw.go.jp/toukei/list/dl/
This study was a joint project of the Japanese Clinical Orthopaedic 20-21-h25.pdf Accessed June 29, 2018.
Association with the Japanese Society for Musculoskeletal Medicine, which 4. Guzman J, Esmail R, Karjalainen K, Malmivaara A, Irvin E, Bombardier C.
contracted with two companies, MINATO Medical Science and OG Wellness Multidisciplinary bio-psycho-social rehabilitation for chronic low back pain.
Technologies, to conduct a clinical trial in order to assess the clinical utility Cochrane Database Syst Rev. 2002;1:CD000963.
of lumbar traction treatment (recipient: Hideki Tanabe). The present 5. Schimmel JJ, de Kleuver M, Horsting PP, Spruit M, Jacobs WC, van Limbeek
biomechanical analysis was conducted as part of that clinical trial. The J. No effect of traction in patients with low back pain: a single Centre,
funding provided by these societies and companies mainly served to cover single blind, randomized controlled trial of intervertebral differential
the fees of disposable research supplies, such as strain gauge or electrodes. dynamics therapy. Eur Spine J. 2009;18(12):1843–50.
6. Wegner I, Widyahening IS, van Tulder MW, Blomberg SE, de Vet HC,
Availability of data and materials Brønfort G, Bouter LM, van der Heijden GJ. Traction for low-back pain with
The corresponding author has full access to all data in the study and or without sciatica. Cochrane Database Syst Rev. 2013;8:CD003010.
assumes final responsibility for the decision to submit this manuscript for 7. Harte AA, Gracey JH, Baxter GD. Current use of lumbar traction in the
publication. The datasets used and/or analyzed during the current study, management of low back pain: results of a survey of physiotherapists in the
which are maintained by designated researchers, are available through the United Kingdom. Arch Phys Med Rehabil. 2005;86(6):1164–9.
corresponding author on reasonable request. 8. Madson TJ, Hollman JH. Lumbar traction for managing low back pain: a survey
of physical therapists in the United States. J Orthop Sports Phys Ther. 2015;
Authors’ contributions 45(8):586–95.
ST organized the biomechanical experiment and conducted the finite element 9. Diab AA, Moustafa IM. The efficacy of lumbar extension traction for sagittal
model simulation. TA, HT, and KF recruited the participants and measured several alignment in mechanical low back pain: a randomized trial. J Back
parameters in their outpatient clinics. TD and MA performed the data analysis. MA Musculoskelet Rehabil. 2013;26(2):213–20.
prepared the first draft of the manuscript, and all authors contributed to writing as 10. Fritz JM, Thackeray A, Childs JD, Brennan GP. A randomized clinical trial of
well as reviewing and approving the final version of the manuscript. the effectiveness of mechanical traction for sub-groups of patients with low
Tadano et al. BMC Musculoskeletal Disorders (2019) 20:155 Page 12 of 12

back pain: study methods and rationale. BMC Musculoskelet Disord.

2010;11:81.
11. Thackeray A, Fritz JM, Childs JD, Brennan GP. The effectiveness of
mechanical traction among subgroups of patients with low back pain and
leg pain: a randomized trial. J Orthop Sports Phys Ther. 2016;46(3):144–54.
12. Shirado O, Doi T, Akai M, Fujino K, Hoshino Y, Iwaya T. An outcome
measure for Japanese people with chronic low back pain: an introduction
and validation study of Japan low Back pain evaluation questionnaire.
Spine. 2007;32(26):3052–9.
13. [No authors listed]. Mechanical lumbar traction. What is its place in clinical
practice? J Orthop Sports Phys Ther. 2016;46(3):155–6.
14. Frost M, Stuckey S, Smalley LA, Dorman G. Reliability of measuring trunk
motions in centimeters. Phys Ther. 1982;62(10):1431–7.
15. Kurutz M. In vivo age- and sex-related creep of human lumbar motion
segments and discs in pure centric tension. J Biomech. 2006;39(7):1180–90.
16. Park WM, Kim K, Kim YH. Biomechanical analysis of two-step traction
therapy in the lumbar spine. Man Ther. 2014;19(6):527–33.
17. Simmerman SM, Sizer PS, Dedrick GS, Apte GG, Brismée JM. Immediate
changes in spinal height and pain after aquatic vertical traction in patients
with persistent low back symptoms: a crossover clinical trial. PM R. 2011;3(5):
447–57.
18. Chow DHK, Yuen EMK, Xiao L, Leung MCP. Mechanical effects of traction on
lumbar intervertebral discs: a magnetic resonance imaging study.
Musculoskelet Sci Pract. 2017;29:78–83.
19. Lai A, Chow DH. Effects of traction on structural properties of degenerated
disc using an in vivo rat-tail model. Spine. 2010;35(14):1339–45.
20. Lee RY, Evans JH. Loads in the lumbar spine during traction therapy. Aust J
Physiother. 2001;47(2):102–8.
21. Cai C, Pua YH, Lim KC. A clinical prediction rule for classifying patients with
low back pain who demonstrate short-term improvement with mechanical
lumbar traction. Eur Spine J. 2009;18(4):554–61.
22. Deyo RA, Battie M, Beurskens AJ, Bombardier C, Croft P, Koes B, Malmivaara
A, Roland M, Von Korff M, Waddell G. Outcome measures for low back pain
research. A proposal for standardized use. Spine. 1998;23(18):2003–13.
23. Suzukamo Y, Fukuhara S, Kikuchi S, Konno S, Roland M, Iwamoto Y,
Nakamura T. Committee on science project, Japanese Orthopaedic
association. Committee on science project, Japanese Orthopaedic
association. Validation of the Japanese version of the Roland-Morris
disability questionnaire. J Orthop Sci. 2003;8(4):543–8.
24. Stokes IA, Gardner-Morse M. A database of lumbar spinal mechanical behavior
for validation of spinal analytical models. J Biomech. 2016;49(5):780–5.
25. Wang S, Wang L, Wang Y, Du C, Zhang M, Fan Y. Biomechanical analysis of
combining head-down tilt traction with vibration for different grades of
degeneration of the lumbar spine. Med Eng Phys. 2017;39:83–93.