Research Review by Dr. Shawn Thistle©


Nov. 2006

Study Title:

Pragmatic randomized trial evaluating the clinical and economic effectiveness of acupuncture for chronic low back pain


Witt CM et al.

Publication Information:

American Journal of Epidemiology 2006; 164(5): 487-496.


Low back pain has a high lifetime prevalence (~70-80%) with an alarming proportion of patients developing chronic pain (up to 7%). Therapeutic management of this condition is varied, and a number of commonly used therapies are still awaiting scientific validation. As more and more patients seek complimentary and alternative treatment methods, it is important to establish which, if any, of these modalities are useful and cost effective in the treatment of low back pain.

Results of previous studies on acupuncture have shown promising, but inconsistent results. Low methodological quality and small sample sizes are consistent downfalls of the existing studies. Further, these studies were mostly experimental, thus providing little information about the application of acupuncture in general medical practice.

This pragmatic study investigated the effectiveness and costs of acupuncture in addition to routine care in the treatment of chronic low back pain. As part of a large research initiative funded by a group of German health insurance funds, this multicentre trial consisted of a randomized and nonrandomized cohort. Patients were included in the study if they received a clinical diagnosis of chronic LBP (> 6 months duration) from a participating physician and either requested acupuncture, or had the physician deem acupuncture an appropriate treatment. All patients were 18 years of age.

Exclusion criteria were standard, and included:
  • protrusion or prolapse of a intervertebral disc with neurological symptoms
  • prior spinal surgery
  • inflammatory, malignant, or autoimmune disease
  • congenital deformation of the spine
  • compression fracture of the spine
  • spinal stenosis, spondylolysis, or spondylolisthesis
This study is unique in that patients had the option to be randomized. Those that did not want to be randomized received acupuncture treatment and were used as a comparison group. Therefore, the three groups involved in this study were:
  1. Acupuncture group (randomized) - 1549 patients who agreed to be randomized received no more than 15 acupuncture treatments in a 3 month treatment period
  2. Control group - 1544 patients who received delayed acupuncture treatment at 3 months with the same restrictions as group 1
  3. Declined randomization acupuncture group - 8537 patients who refused randomization and received similar treatment to group 1
The treatment phase was three months in length with a three month follow-up period (therefore six months total for each patient). Acupuncture points utilized and number of needles was left to the discretion of the treating physician (hence the "pragmatic" label on the study). Only needle acupuncture was allowed, and all patients were allowed to utilize conventional treatments as needed.

Unfortunately no details were given on the "conventional treatments" patients received or any confounding effects they may have had on the results.

The primary outcome measure in this study was back function at 3 and 6 months, assessed by the Hannover Functional Ability Questionnaire (previously validated). Patients who demonstrated at least a 20% improvement in "back function loss" were considered to be successful cases.

The cost perspective was evaluated at a societal level. Costs for physician or hospital visits, medication, acupuncture or other treatments, and number of sick leave days were provided by the health insurance providers.

Statistical analysis then generated a Quality Adjusted Life-Years cost (QALY) similar to what other large scale studies have done in the past (the UK BEAM low back pain study comes to mind here - reviewed in the Chiropractic/manual Medicine section of the database).

Pertinent Results:

  • the randomized groups were comparable in all baseline characteristics
  • patients who refused randomization were more likely to have greater than 10 years of schooling and have more severe pain, but they had a shorter duration of symptoms by about 1 year compared to the other groups
  • in the first three months, patients in the acupuncture groups received an average of 10.4 treatments
  • back function improvement was more pronounced in the randomized acupuncture group compared to the control group
  • the proportion of responders (successful treatment outcome) was 52.6% in the acupuncture group versus 26.8% in the control group (remember they received "usual care")
  • in the non-randomized patients, back function improvement was greater than the randomized acupuncture group, as well as pain and quality of life measures (both secondary outcome measures)
  • acupuncture's effects on back pain and function were more pronounced overall in patients with more severe initial presentation, and younger patients
  • other factors that contributed to favorable outcome included having more education, higher mental quality of life, and greater experience in acupuncture treatment
  • longer term follow up at six months revealed only slight reductions in treatment effect in all acupuncture patients
  • patients who received delayed acupuncture (the control group) demonstrated equivalent improvement to the other two acupuncture groups

Conclusions & Practical Application:

Patients with chronic low back pain treated with acupuncture in addition to routine care showed significant improvement in symptoms and quality of life compared to those who did not receive acupuncture. Acupuncture was associated with higher overall cost, but was estimated to be cost effective based on the observed improvements.

I think pragmatic trials have an important place in evaluating CAM therapies. What better way to evaluate a therapy than to allow practitioners to apply it within the context of their normal daily practices? Some methodological shortcomings are unavoidable in this type of study. Blinding of patients and doctors was not possible, and when other therapies are allowed, confounding may occur.

This study is one of the largest acupuncture studies done to date, and provides further evidence that acupuncture is an effective adjunct therapy in the treatment of chronic low back pain.

I find it interesting that two factors were associated with better response to acupuncture - namely younger age and increased severity of pain. Anecdotally, this mirrors trends I have noticed in my practice. Younger patients with severe pain tend to respond favorably to acupuncture.

Perhaps future research will help us identify those most likely to respond to acupuncture, similar to the clinical prediction rule currently in development for spinal manipulation in LBP.