The purpose of this blog is to review the literature on lumbar traction and its effect on pain, function, and disability. We will discuss the different types of lumbar traction, the theories behind its use, and its effects on low back pain, function, and disability, and adverse events.
What is lumbar traction?
Lumbar traction is a modality in which are forces applied in an attempt to create some form of separation between lumbar vertebrae. Force is generally applied in a superior-inferior direction, with one side typically “fixed” in place while the force is applied to the other side.
The earliest historical records of spinal traction being utilized for treatment are from 1800 BC, and by 500 BC Hippocrates created the first known spinal traction device. In the 1950s James Cyriax, one of the earliest orthopedic rehabilitation practitioners, promoted lumbar traction for treatment of various back issues – particularly lumbar disc lesions. Since that time, devices and treatments have progressed in complexity and theorized specificity.
What are the categories of lumbar traction?
There are 37 main types of lumbar traction reported in the literature. These types can be divided into different divisions across 3 main categories: duration, direction of force, and type of force application.
There are three subcategories of duration – continuous, sustained, and intermittent.
- Continuous traction
- This is generally from multiple hours to days.
- Sustained traction
- This is generally 20-60 minutes.
- Intermittent traction
- This is generally a few seconds to a few minutes with multiple cycles.
Direction of Force
There are three subcategories within the direction of force category – axial, positional and distraction-manipulation.
- A longitudinal force is applied and the patient is typically able to adjust and regulate the force (for example gravity assisted or pool traction).
- This may be off-axis, and the patient or clinician may vary the direction.
- This is a longitudinal force in which the clinician determines the magnitude and direction.
Type of Force Application
There are two main types of force application done with traction – either mechanical or manual.
- This is the most popular and well-researched form of traction. In this type of traction, the machine provides the force which is controlled and regulated. We will focus on this method moving forward as there is more robust research on it than manual traction.
- Manual traction is commonly used by clinicians, and it is claimed that it can deliver a more individualized dosage to specific, isolated joints. We will only discuss this type of traction briefly as it is not as well-researched as mechanical traction due to control limitations.
What are the parameters for traction?
Across each of the categories there is a huge range of parameters that are used. The duration varies from a few minutes to many hours. The force applied will range from a few pounds to 50% of bodyweight. The cycle time will vary from a few seconds, to minutes, to none at all. The rest duration will fluctuate from none to minutes.
It appears that the majority of parameters were decided quite arbitrarily as there is little to no justification for them provided in the research. Airwaily et al. report that there is a challenge in suggesting any dosage or parameters based on the evidence as there is little consistency per condition or type of traction. As we will demonstrate in future sections, there does not seem to be a significant difference in effects between the various parameters.
What is the theory of lumbar traction?
There have been several theories on the various benefits of traction dating as far back as ancient times. There are several theories across different beliefs on the mechanism of traction from a mechanical perspective:
- Increasing in intervertebral space
- Increasing in intervertebral foramen
- Greater tension across the posterior longitudinal ligament
- Reduction in lumbar lordosis
- Suction of a disc herniation back to disc space
Due to these various mechanisms theorized, there are different forms of traction with different protocols claimed to provide a range of effects – some arguing superiority over others. We have pulled in the best studies on these topics to examine the theories and their effects.
What are the treatment effects?
Up to this point, we’ve focused on general knowledge of the topic. Now we are going to be more critical and analytical of the evidence. In considering the treatment effects, there is a range of information that is shared from studies, typically looking at outcomes across anatomical changes, pain changes, and functional/disability changes. As we go forward, it is important that we consider them both separately and collectively.
There are a few different anatomical effects claimed by early proponents of modern spinal traction methods. A main one was the retraction of herniated material from the spinal discs. Onel et al 1988 examined this with a high load spinal traction to patients with recent lumbar disc herniations and did find that there was a reduction in herniated disc material – ranging from 57-78% depending on the location of the herniation. Ozturk et al. 2006 had similar findings when examining spinal traction with continuous parameters for patients experiencing back pain and sciatica.
The Ozturk study differed though as it also had a control group. Now the traction group did in fact have a greater reduction in herniation size than the control group, but both groups had improvements across the short time span of the study. This study did not examine effects on pain or function.
Kamanli et al. 2010 conducted a similar study, except this time it was with the use of intermittent traction where the force was provided for 10s, then relaxed for 10s, and repeated. Similar to the other studies, there was a reduction in disc herniation size in 5 of the 26 patients, however there was no significant change in 18 patients, and 3 patients actually had an increase in disc herniation size.
Other anatomical theories on the effects of traction are on the gapping effects and tensioning of the ligaments of the spine. Twomey 1985 examined this and found that when providing a sustained traction to cadavers, there was a creep effect, providing tensioning to the ligaments.
Janke et al. 1997 had patients perform gravity dependent traction with a device to put them in an inverted position to their control. The authors then measured changes in the lumbar spine and found lumbar lengthening and spinal curvature degree changed with the intervention.
Kane et al. 1985 found that with gravity dependent traction there was a significant separation between lumbar segments after 10 minutes of time spent dependent. There was no control group or variation to compare to, so we have this limited information to go with.
Most of the studies reviewed found these differing changes, which are assumed to be benefits (we will see in the next section whether these show positive benefits or not), but did not examine if they are retained. Santos et al. 2010 compared two different traction protocols – one using 10% of bodyweight and one using 50% of bodyweight. The authors found that stature was increased immediately following traction, which was presumably due to increased spacing at the intervertebral discs, and that this effect was more prominent in the 50% bodyweight condition than in the 10% bodyweight condition. However, Santos followed up testing just 10 minutes later and saw a decrease to almost baseline.
Pain & Disability Effects
Traditional beliefs on pain were heavily focused on a direct link to the anatomical state of the human – which we’ve learned is not accurate. In our previous section, we saw that there is a mix of information, but a general trend in favor of anatomical effects that should be “beneficial.” However, as we’ve seen in other literature on pain and anatomy – there is not a direct link or causation.
Beattie et al. 2008 published a study examining 296 patients who had low back pain with evidence of disc degeneration and/or a herniated disc at 1 or more levels in the lumbar spine. The patients received 8 weeks of lumbar traction consisting of 30 minute sessions, 5 times a week for the first 4 weeks and then 1 time a week for the second 4 weeks. The patients reported significant improvements in both pain and disability at a 30 day and 180 day follow up. Unfortunately this study has some major limitations – particularly that it did not have any control group.
Pal et al. 1986 looked at a trial of continuous lumbar traction for those experiencing back pain and sciatica and did find favorable results for the treatment. However, the control group had the same results as the traction group and the authors stated:
“The findings of this study question the justification of admitting patients with back pain into hospitals for purposes of traction.”
Similarly, Matthews & Hickling 1975 completed a double blinded study examining the effects of sustained lumbar traction on a group of patients with back pain and sciatica compared to a control group. Both groups received treatment 5 times a week for 3 weeks, but the traction applied to the control group did not have enough force to overcome friction. There was a trend towards improvements in pain and a straight leg raise test in both groups, and no statistically significant difference between groups.
In the prior section on changes in anatomical factors, we discussed the Kamanli et al. 2010 paper which found individuals had a decrease, no change, or an increase in disc herniation size following 15 treatments of intermittent traction. Fascinatingly, the patients saw improvements in pain, reduction in sleep disturbances, and increase in global assessments – regardless of anatomical changes.
Ljunggren et al. 1992. Performed a study examining the effects of manual traction vs isometric exercises on patients with herniated lumbar discs. This study found both groups improved at a similar pace and had no significant difference.
Interestingly, Borman et al. 2003 ran a study comparing physical therapy and physical therapy with the addition of sustained traction. At the end of the study, both groups had significant changes for disability, pain, and satisfaction rates, with maintained results at a 3 month follow up. Most of the time when we see a study that does A vs A + B, the latter is favored to have a greater effect.
Interestingly, the group that did not receive the additional intervention of traction had an overall greater outcome than the physical therapy + traction group. Furthermore, the patients who reportedly continued exercises at home from time of study completion to the follow up (51% of patients) had a greater reduction in disability than those who did not continue their exercises
Similarly to the Borman et al. 2003 study, Rattantham et al. 2004, Harte et al. 2007, Beurskens et al. 1997, Beurskens et al. 1995, and Schimmel et al. 2009 conducted similar studies of “regular” physiotherapy and physiotherapy plus traction. These studies also found no difference in pain or disability after weeks of treatment.
Unlu et al. 2008 studied a comparison on effects of traction, ultrasound and low-power laser for the treatment of acute leg pain and low back pain after confirmation of a disc herniation. All groups had significant reductions over 3 weeks of treatment, with no difference between the groups.
When is traction contraindicated?
There are various reasons traction is contraindicated with the main reasons being spinal compromise (malignancy, infection, osteoporosis, inflammatory condition, fracture), systemic risk (hernia, pregnancy, severe hypertension, severe cardiovascular or respiratory disease), and fear of traction.
There is risk with any treatment. We always need to deem if the benefit outweighs the risk of harm. In the case of traction, there are some considerations that need to be addressed.
Our first one is from the points discussed prior in the Kamanli et al. 2010 paper in which we saw a similar number of patients had an increase in disc herniation size as those who had a reduction in size. If we practiced on a biomedical model with an emphasis on anatomical changes for treatment effects, this would be a serious concern.
Many might think this is a one-off study, but we actually have more information reaffirming this is possible – regardless of the suction theory. Deen Jr, et al. 2003 report a case in which an individual who was performing vertebral axial decompression traction therapy had experienced sudden, severe radicular pain and required immediate surgical intervention. Further, a case study from Kim et al. 1999 report a patient who sustained a lumbar disc herniation and cauda equina syndrome following a self traction treatment, resulting in serious symptoms and requiring immediate surgery.
Wegner et al. conducted a systematic review of the randomized control trials on traction and found that the majority of studies did not report whether or not adverse events occurred. In the studies that did report them, there were adverse events of increased pain, aggravation of neurological factors, and surgery following traction. Most of the studies on traction are of moderate to low quality with a moderate to high risk of bias, which means there are likely many adverse events not being reported.
Cumulative consideration – Where does all of that information guide us?
Madson et al. 2015 surveyed physical therapists in the US and found that 58% used traction therapy for nerve root compression. This is a very high rate, which should be supported by quality evidence demonstrating efficacy in the intervention.
We’ve reviewed numerous papers that highlight a few key things:
- Traction can have a mechanical and anatomical effect.
- It may be superior to placebo, but it may be equivalent or inferior to it as well.
Krause et al. 2000 reviewed literature on the intervention and found that if traction were to be of any benefit it should be applied in the acute time frame. The authors report that there isn’t a dose response relationship, so a low dose treatment should be more than sufficient.
In trying to determine who is most likely to respond to traction, Cai et al. 2009 made a clinical prediction rule which gave a positive likelihood ratio of 9.36 for increased response rate to traction. This rule was centered on four variables
- Non-involvement in manual work as a profession
- Low level fear avoidance beliefs
- No neurological deficit
- Above 30 years of age
Why these four things? We are not sure and we could postulate but the best thing we can take away from it is if your client doesn’t match up to these things, traction is likely not going to be beneficial at all.
An interesting result from the Twomey 1985 study we mentioned in the prior section was that the greatest amount of separation occurred at the disc levels with the least amount of disc degeneration. If we were practicing through a biomedical model, we would want to see the distraction effect occur at the levels with degeneration, but not the other levels. This result provides an unique challenge to that working model