
Having recently stumbled upon a discussion on a “quackbusting” website mocking chiropractic’s warning about NSAID use due to the associated dangers, it reminded me of a piece i wrote a while back comparing the evidence for NSAID use and spinal manipulation for low back pain. Below is the section on systematic reviews:
None of these systematic reviews looked exclusively at direct comparisons of spinal manipulation and NSAID use in the treatment of back pain. Instead they compared a range of treatment modalities including, manipulation and NSAIDs, in the treatment of back pain.
Ferriera et al (2002) is a broad systematic review of RCTs, looking at the efficacy of spinal manipulation compared to many other treatment methods for chronic low back pain (LBP>3 months). “Spinal manipulation”, in this trial, included a broad range of techniques, not only HVLA thrusts. The review’s aims were clearly stated and the search methods were rigorously applied. 2 independent reviewers screened and selected the trials to be included. The PEDro (1999) scale was used to assess trial quality, with those scoring less than 3 out of 10 excluded.
The main results were that; no significant difference was found when comparing spinal manipulation to NSAIDs in the treatment of chronic low back pain. A 95% confidence interval was calculated for between-group differences to determine the treatment effect for each outcome. A relative risk of 0.7 or less was stated to be clinically significant. This was a good quality review although the literature directly comparing manipulation and NSAIDs to each other was very limited (Bronfort et al 1996; Giles & Muller 1999). Only 8 out of 117 RCTs were included which limits the scope of the review and only 5 trials scored more than 5 out of 10 on the PEDro scale with the Giles & Muller (1999) trial scoring only 4.
A systematic review by Van Tulder et al (1997) and a more recent systematic review, of systematic reviews, by the same authors (Van Tulder et al 2005) looked at the effectiveness of the most common types of treatment for patients with non-specific low back pain (LBP). These included patients with acute (<6weeks) and chronic (>12 weeks) LBP and studies involving at least one of several other treatment interventions, as well as placebo and sham treatment comparisons. These were much larger systematic reviews than the Ferriera et al (2002) trial, including 150 RCTs in the Van Tulder et al (1997) review. Both reviews had clearly expressed aims, and rigorously applied search methods. In the Van Tulder et al (1997) trial, 2 independent reviewers assessed each study, out of 100, based on previously accepted criteria. A study was judged to be of high quality if it scored more than 50 points. Only 59% of the RCTs were found to be of high quality. Two independent reviewers in the Van Tulder et al (2005) review assessed trial quality using the Cochrane Back Review Group (CBRG) criteria. A trial was judged to be of high quality if it scored 6 points or more out of 11 on the CBRG scale.
There was no discussion about the self-limiting nature of acute LBP, in these studies, and the clinical relevance of its inclusion must be questioned (Cedraschi et al 1999; Waddell 2004). For this reason the fact that Van Tulder et al (1997) found that for acute LBP, NSAIDs were more effective than placebo whereas manipulation was not, is not clinically relevant. This also applies to the findings by Van Tulder et al (2005) that NSAIDs were significantly better than placebo for acute LBP and that manipulation had statistically significant short term improvements in pain compared to sham therapy for acute LBP (Van Tulder et al 2005).
Both these systematic reviews included the use of either a placebo or sham treatment, which is inappropriate when assessing manipulation. It is not clear how a practitioner would apply a placebo/sham manipulation, nor is it clear whether a patient would believe the treatment was genuine (Andersson et al 1999). In a recent study by Hancock et al (2006) to select an appropriate placebo to use in RCTs assessing spinal manipulation, there was widespread disagreement amongst the experts as to the suitability of placebo use. For this reason the findings of Van Tulder et al (1997), that manipulation was more effective than placebo for chronic LBP, are of little clinical relevance. This also applies to the claim that, manipulation was statistically significant in its effective reduction for short and long term pain relief and improved function compared to sham treatment for chronic LBP (Van Tulder 2005). The methodology of both reviews by Van Tulder et al (1997 & 2005), was generally appropriate, however because of the poor clinical relevance of the studies outlined above, it is difficult to apply any of the findings to clinical use.
There were a limited number of relevant trials comparing these 2 interventions for back pain and a clear lack of high quality reviews which carried enough methodological weight to change current practice. Many of the weaknesses were common to all the studies e.g. lack of blinding of patients and practitioners, small group numbers, multiple osteopaths/medics providing treatment and the potential for non-disclosure of self-medicating/exercise by patients. Most trials found no significant difference between the two types of treatment and those that did suffered from methodological weaknesses which reduced their quality. These methodological weaknesses might have been reduced if qualitative instead of quantitative research were carried out.
In looking at spinal manipulation in isolation from its usual therapeutic context, the clinical relevance is reduced. However to compare a multi faceted osteopathic treatment with other interventions would fail to isolate the effective components of the osteopathic consultation.Given the well documented risks associated with NSAID use and since NSAID is evidenly no more effective than manipulation for LBP, i sugggest spinal manipulation is the safer and therefore preferable option.
Refs:
Andersson, G; Lucente, T; Davis, A; Kappler, R; Lipton, J; Leurgans, S. (1999). A comparison of osteopathic spinal manipulation with standard care for patients with low back pain. The New England Journal of Medicine, Volume 341(19), pp 1426-1431.
Bronfort G. Goldsmith C. Nelson C. Boline P. Anderson A. (1996). Trunk exercise combined with spinal manipulative or NSAID therapy for chronic low back pain: a randomized, observer-blinded clinical trial. Journal of Manipulative & Physiological Therapeutics. 19(9):570-82.
Cedraschi C, Robert J, Goerg D, Perrin E, Fischer W, Vischer T. (1999) Is chronic non-specific low back pain chronic? Definitions of a problem and problems of a definition. British Journal of General Practice, 49(442):358-62.
Ferreira M, Ferreira P, Latimer J, Herbert R, Maher C. (2002) Does spinal manipulative therapy help people with chronic low back pain? Australian Journal of Physiotherapy Vol. 48. pp 277-284.
Giles L, Muller R. (1999) Chronic spinal pain syndromes: A clinical pilot trial comparing acupuncture, a non-steroidal anti-inflammatory drug, and spinal manipulation. J Manipulative Physiological Therapeutics. 22:376–81.
Hancock M, Maher C, Latimer J, McAuley J. (2006) Selecting an appropriate placebo for a trial of spinal manipulative therapy. Australian Journal of Physiotherapy. 52(2):1358.
Van Tulder M, Koes B, Bouter L. (1997) Conservative treatment of acute and chronic non-specific low back pain. A systematic review of randomized controlled trials of the most common interventions. Spine. 15; 22(18):2128-56.
Van Tulder M, Koes B, Malmivaara A. (2005) Outcome of non-invasive treatment modalities on back pain: an evidence-based review. European Spine Journal. Sup1: 64-81.
Waddell G. (2004) The Back Pain Revolution. 2nd Edition, Churchill Livingstone. p.33.