Spinal manipulation for back and neck pain has been and remains controversial. There have been numerous randomized trials, meta-analyses, and reviews with no consensus, and given the nature of the entrenched positions on both sides and methodological problems with available studies, there is not likely to be a resolution in the near future.
The reported rationale for spinal manipulation in the treatment of back and neck pain is variable and ranges from "reduction of a bulging disc, correction of the internal displacement of disc fragments, and freeing of adhesions around a prolapsed disc or facet joints to inhibition of transmission of nociceptive impulses" (Koes et al 1991, p. 1298).
In many studies the specific manipulation or mobilization techniques are inadequately specified. Manipulation involves a "high velocity thrust to a joint beyond its restricted range of movement," whereas mobilization uses "low velocity passive movements within or at the limit of joint range" (Koes et al 1991, p. 1298), although many studies and reviews use the term "manipulation" to encompass both manipulation and mobilization (Koes et al 1991).
Various systematic reviews have noted variable but generally poor methodological quality among existing clinical trials (Koes et al 1991; 1995; 1996; Ernst 2003; Gross et al 2010) or reviews (Assendelft et al 1995) of spinal manipulation. Identified methodological problems of published trials include small sample sizes, use of nonhomogenous study populations, failure to describe the manual procedure used, no description of the randomization procedure, no detailed tallies or description of study dropouts, absence of a placebo-control group, lack of blinded outcome assessments, and lack of any long-term outcome data (Koes et al 1995; 1996; Barr 1996; Hurwitz et al 1996). Methodological quality has been associated with the outcomes of available clinical trials, with better-quality trials less likely to report benefit (Koes et al 1995). Moreover, the presence of a spinal manipulator in the review team was associated with a positive conclusion of a published review, as determined in an analysis of the methodological quality and conclusions of published reviews of spinal manipulation; this suggests a component of reviewer bias in some published reviews (Assendelft et al 1995).
Several systematic reviews have concluded that spinal manipulation and mobilization probably provide short-term benefits for some patients with neck pain (Hurwitz et al 1996; Gross et al 2010) or low-back pain (Hoehler et al 1981; Abenhaim ad Bergeron 1992; Shekelle et al 1992; Powell et al 1993; Assendelft et al 1995), particularly for those with uncomplicated acute low-back pain (Shekelle et al 1992), but "strong conclusions are precluded" by the overall poor quality of available trials and reviews (Assendelft et al 1995; Gross et al 2010). Koes and colleagues concluded more strongly that the "efficacy of spinal manipulation for patients with acute or chronic low back pain has not been demonstrated with sound randomized clinical trials" (Koes et al 1996, p. 2860). No convincing data is available to indicate any relative benefit of spinal manipulation compared with other conservative therapy for acute neck or back pain or any long-term benefit for neck or back pain (Hoehler et al 1981; Shekelle et al 1992; Powell et al 1993). Similarly, Ernst concluded that "the notion that CSM [chiropractic spinal manipulation] is more effective than conventional exercise treatment in the treatment of neck pain was not supported by rigorous trial data" (Ernst 2003).
The complication rate is unknown but generally felt to be low (though estimates are highly variable and based on limited data on procedure frequency and usually literature case reports or administrative data, with many untested assumptions) (Shekelle et al 1992; Hurwitz et al 1996). Nevertheless, a systematic review conducted by the RAND Corporation, the UCLA Department of Epidemiology, the West Los Angeles Veterans Affairs Medical Center, and various chiropractic organizations concluded that "although the complication rate of manipulation is small, the potential for adverse outcomes must be considered because of the possibility of permanent impairment or death" (Hurwitz et al 1996).
Because the efficacy of manipulation is apparently no better than mobilization, and because manipulation is more likely to be associated with serious complications, especially when applied to the cervical spine, some have advocated that the use of manipulation with high-velocity thrusts be abandoned for the treatment of patients with neck pain (Barr 1996). Similarly, because there appear to be common features among cases with serious neurologic complications reported in the literature (eg, manipulation performed in the presence of a herniated nucleus pulposus or manipulation of the cervical spine), Powell and colleagues concluded that the risk-benefit ratio was unacceptably high for patients with "radicular symptoms or signs associated with prolapsed discs and neck pain" (Powell et al 1993). The authors added that "potential complications and unknown benefits indicate that [spinal manipulation therapy] should not be used in the pediatric population."
Ernst reviewed all 7 sham-controlled trials of spinal manipulation and found them to be of variable methodological quality, but collectively they did not show therapeutic effects beyond placebo, and the 3 most rigorous trials were negative (Ernst 2000). Ernst concluded that the "results available to date suggest that the therapeutic success of spinal manipulation is largely due to a placebo effect."
In conclusion, existing clinical trials and reviews of spinal manipulation have been of variable but generally poor methodological quality. The limited available data suggest a short-term benefit of spinal manipulation for uncomplicated (nonradicular) back and neck pain, but the generally low methodological quality of available studies precludes a firm assessment of the benefit of spinal manipulation. No convincing data is available to indicate any relative benefit of spinal manipulation compared with other conservative therapy for acute neck or back pain, or any long-term benefit for neck or back pain. Uncommonly spinal manipulation is associated with severe and sometimes fatal neurologic complications.