Research Review By Dr. Rob Rodine©

Date Posted:

July 2009

Study Title:

A clinical prediction rule for classifying patients with low back pain who demonstrate short-term improvement with mechanical traction

Authors:

Cai C, Hao Pua, Y & Chong Lim K

Author's Affiliations:

Rehabilitation Department, Alexandra Hospital, Singapore.

Publication Information:

European Spine Journal 2009; 18: 554-561.

Background Information:

Dr. Shawn Thistle recently reviewed the article titled ‘’Development of a clinical prediction rule to identify patients with neck pain likely to benefit from cervical traction and exercise.’’ In this review, Dr. Thistle identifies how the proposed clinical prediction rule (CPR) may aid clinicians in determining which patients are most suited for cervical traction.

A similarly proposed CPR, this time in the low back, is reviewed here from a group of researchers in Singapore. The authors highlight the high lifetime prevalence of low back pain (80%) and the associated high economic burden. However, more worrisome is the confusion surrounding the efficacy of numerous conservative therapies, including traction.

Previous literature has pointed towards a lack of evidence for traction in low back pain patients. However, there is strong concern regarding the quality of the studies, the lack of consistency between studies and the small sample sizes. In addition, studies have not focused on pre-treatment fear-avoidance status, as pointed out by the current authors.

The purpose of the current study was to identify the correct subgroup of low back patients that are most likely to benefit from lumbar spine mechanical traction. This is done through the identification of variables which may have a significant relationship with patient outcome following a course of treatment. Therefore, if these variables are found during patient screening it may better inform prognosis as well as direct specific treatment.

Pertinent Results:

If you are unfamiliar with the basic setup for early studies on clinical prediction rules, please refer first to the “Study Methods” section below before proceeding with the results of this study listed here:
  • 13 potential prediction variables were identified. These included age, BMI, mechanism of onset, prior history of low back pain, presence of pain below the knee, job status, standing as an aggravating position, sitting as a relieving position, right SLR, neurological deficit and FABQ score.
  • When comparing the responders to the non-responders, there was a statistical significant difference regarding age, BMI, having a traumatic onset and pain below the knee (both higher in non-responders), manual and desk work, standing as an aggravating position and sitting as a relieving position, the right SLR, neurological deficit (more non-responders) and a higher work subscale score on the FABQ (for non-responders).
  • When the predicting factors were identified and entered into forward stepwise logistic regression, having an FABQ score of less than 21, not demonstrating neurological deficit, having an age older than 30 and no involvement with manual labour were identified as the four variables which composed the CPR.
  • Of the 25 subjects who were deemed to be responders to mechanical traction, 9 had all four predictors, 19 had three or more, 24 had two or more and 25 had one or more.
  • Patients without neurological deficit were 12.75 times more likely to respond to traction than their counterparts.
  • Patients with a lower FABQ score (<21) were 3.07 times more likely to respond to traction than their counterparts.
  • Patients that were not involved in manual labour were 3.66 times more likely to respond to traction than their counterparts.
  • Patients over the age of 30 years were 4.18 times more likely to respond to traction than their counterparts
  • If 3 of 4 variables were present in the patient, the probability of successful traction was deemed to be 42.2%.
  • If 4/4 variables were present in the patient, the probability of successful traction was deemed to be 69.2%.

Clinical Application & Conclusions:

While this study demonstrates several limitations, it offers and excellent preliminary attempt at identifying variables in low back pain patients most likely to respond to mechanical lumbar traction. Four variables emerged from the analysis in this study as helpful in predicting a positive response to traction:
  1. being > 30 years of age
  2. having no neurological deficit (this contradicts previous findings from other low back pain CPR research: see review titled “Low Back Pain – Clinical Prediction Rule” below)
  3. having a low FABQ score
  4. non-involvement in manual labour
The authors caution however, that this CPR should not be used in a multi-modality treatment regime, based on the isolated intervention applied during this study. However, the authors also caution that this study should not be used to formulate treatment strategies longer than 3 sessions over a 9 day period.

Regardless, this study is useful in assisting clinicians to identify patients most suitable to traction and allow further prognostication surrounding low back pain patients presenting for conservative therapy. This is helpful when discussing goals and objectives from treatment as well as comparing treatment options. Though there is an obvious need for further research to confirm, this is an excellent first step.

Study Methods:

The current study was based on 129 patients that were referred from the orthopedic outpatient clinic over a 6 month time period with a chief complaint of pain and/or numbness in the lumbar spine, buttock and/or lower extremity. Patients were excluded based on current pregnancy, signs of spinal cord injury, history of spinal fracture, surgery or osteoporosis, or if it was determined they had non-spinal pain.

Outcome measures used and collected prior to treatment included the modified Oswestry Back Disability Questionnaire (OBDI), the Fear Avoidance Beliefs Questionnaire (FABQ), the numerical pain rating scale (NPRS), active lumbar flexion (recorded based on finger tip position at end range to mid-thigh, patellar, mid-shin and distal-shin) and the straight leg raise (SLR) determined at limitation caused by pain. In addition, standard physical examination findings such as neurological status and segmental mobility testing were conducted.

The intervention provided to patients was three motorized lumbar traction sessions over a 9 day period, using the Triton DTS Traction System from Chattanooga. Flexion sensitive patients were positioned supine with the hips and knees flexed to 90 degrees, and the legs supported on a stool while traction was applied. Patients that were not sensitive to flexion received traction in the supine position. Applied traction force was judged to be 30-40% of the patient’s body weight, intermittently at 30 seconds on, with 10 seconds off for 15 minutes.

Following the trial of therapy, patients were provided with the OBDI once more. Those patients with an OBDI score improving 50% or more were deemed to be responders to treatment. Data belonging to responders was then compared to non-responders in order to determine possible predictive variables.

Of the 129 recruited patients, all were included. Twenty five patients (19.4%) were found to have a 50% improvement score following 3 sessions of mechanical traction, and termed the responders.

Study Strengths / Weaknesses:

This study has a few major limitations which are important to address. First is the sample size. The authors are comparing a group of only 25 responders to that of 104 non-responders. A larger sample size yielding a greater number of responders would provide greater strength to these conclusions. In addition, only one therapy was utilized without a control group. Therefore, it is hard to determine if the 25 responders would have responded equally to another form of therapy or if their improvement was due to natural history over the course of the 9 days. In addition, the sample used was quite heterogeneous for gender (83.7% male) and duration of pain (ranging from 1-1,040 weeks).

Furthermore, the use of an OBDI improvement of 50% or more was not explained adequately. One can wonder how the results of this study would change if an alternate improvement score were chosen. For example, the OBDI has been demonstrated to show minimum clinically important difference at a change in score ranging from 6 to 15 points (12-30%). Despite the authors explaining this choice based on the ‘mechanical characteristics of traction’ being utilized as an adjunct treatment to other interventions, a more descript rationale is required given their methodology.

Additional References:

  1. Davidson, M et al. A comparison of five low back disability questionnaires: reliability and responsiveness. Phys Ther, 2002, 82(1):8-14.
  2. Fritz JM et al. A comparison of a modified Oswestry low back pain disability questionnaire and the Quebec Back Pain Disability Scale. Phys Ther, 2001, 81(2): 141-144.