Acupuncture as a possible treatment for chronic lower back pain
A position paper for IPHY undergrad students
Ellora Weston and Krista Miller
Spring 2005
A paper for The Clinical Journal of Pain
Chronic lower back pain (cLBP), defined as pain lasting more than 12 weeks (van Tulder et al., 1999), has become a considerable medical issue in Western society, which has encouraged researchers to discover a long-term effective modality to treat patients suffering from this condition. Recently, acupuncture has been in the spotlight in Western cultures as a possible treatment. The primary question in the current literature is “How effective is acupuncture as a treatment for adult men and women suffering from cLBP in Western civilization?”
This paper is a review of 10 studies (8 research reports and 2 position papers) performed from 1980-2003. All research studies were cross-sectional (Leibing et al., 2002, also had a longitudinal component) with varying levels of blinding and randomization, and subjects were all adult men and women suffering from cLBP that is not better explained by a medical disorder. Results varied between finding significant benefits to no significant changes in pain levels after acupuncture treatment.
Originating in China, the theory of traditional acupuncture relies on the body’s balance of the inner yin and yang forces. Furthermore, Chinese doctors believe that most medical complaints can be alleviated by activating a combination of 361 specific points on or just below the surface of the skin (van Tulder et al., 1999). These points are a connection of complex energy lines that are said to stabilize the yin and yang forces. To date, the effectiveness of acupuncture for treating cLBP on Western patients has not yet been scientifically established. The current understanding, according to the National Institute of Health, is that while acupuncture is widely used and patient testimonials are often positive, “the issue is complicated by inherent difficulties in the use of appropriate controls, such as placebos and sham acupuncture groups” (Acupuncture, 1997). However, studies with large test groups, longitudinal methods, double-blinded and randomized controlled designs are beginning to further the Western knowledge base of this ancient tradition. Results have yet to be conclusive (see Table 1 and Figure 1). Several studies are in agreement that acupuncture is, indeed, an effective method of treatment for cLBP, whereas others are equally in agreement that the results are solely due to poorly designed studies and the placebo effect.
Acupuncture is recently a topic of consideration among cLBP patients who have not received gains from several Western treatments, including conventional orthopedic treatment (COT), such as spinal manipulation, pain medication, traction, and anti-inflammation procedures. Despite preconceived notions about needles, acupuncture is relatively non-invasive. Although in Western cultures acupuncture is approached as an adjunctive treatment to other forms of non-invasive treatments like spinal manipulation, the studies in this review examined acupuncture as an isolated treatment.
This paper began as an analysis to specifically address the effectiveness of acupuncture exclusively in Western cultures for the purposes of further educating upper division IPHY students who may be potential future researchers in the field. However, after reviewing the literature to date, we came to the conclusion that many of the studies performed on both sides of the argument are not methodologically reliable. Therefore this paper focuses more on an analysis on the literature, arguing that acupuncture is overall not effective due to the extensive limitations found in the studies supporting acupuncture as a viable treatment.
We assume that the bias in Eastern culture could lead to positive results of pain alleviation. In Western cultures, acupuncture is not established as effective or ineffective, whereas in China, acupuncture is frequently sought out as a first option for treatment. Because cLBP patients in Western cultures have not been raised with a belief either in favor or against acupuncture, we speculate that studies performed on Western subjects by highly trained acupuncturists will have a smaller probability of results due to the placebo effect than studies performed in Eastern cultures.
Several recent studies have found significant decreases in pain intensity of patients suffering from cLBP, using the Visual Analog Scale (VAS) scale. The VAS is a self-reported pain scale and was the primary outcome measure in most of the studies reviewed. On the other hand, there are several researchers that have found no benefit from acupuncture. This paper will analyze the discrepancies in the research and come to a better understanding of the research that has been performed to date.
Despite the growing popularity of acupuncture as an alternative treatment to conventional Western methods, it appears that the most credible studies in the current body of knowledge agree that acupuncture does not significantly improve pain in patients who are suffering from cLBP. Strong evidence from researchers also supports the conclusion that acupuncture is not more effective than conventional Western approaches. Additionally, it has also been suggested that the placebo effect largely explains any apparent effectiveness of acupuncture in terms of cLBP management.
After reviewing the studies to date, we have concluded that there is not sufficient convincing evidence in favor of acupuncture’s effectiveness as a treatment for cLBP. Although all of the reviewed studies do not contain the required criteria for a strong, powerful study in this field (see Figure 2), we believe that the studies finding no significant benefit from acupuncture were overall performed with greater internal validity. The research that is available in the current body of knowledge is inherently flawed because of a variety of internal study limitations. Due to the low credibility of the results that have been found, this review is not focused on which side of the argument has greater strengths. Rather, we have examined the studies to find which side contains the lesser of the limitations in terms of severity of effects they may have had on the results reported. The primary limitations we found were: inconsistent treatment protocols, performing the acupuncture treatments with different acupuncturists of unknown or unstated experience, not stating the extent (if any) of blinding in the study design, and small treatment group sizes. These limitations cannot be overlooked. It is difficult to speculate how each study’s results might have changed had they been stronger and more internally reliable. We examined each limitation individually to determine how much of an impact it may have had on the results presented. They are presented in order where the highest ranked is most likely to dramatically influence results and the lowest is least likely to do so. By doing this, it is apparent that the studies with the least severe limitations consistently found acupuncture to be an ineffective or less effective treatment for cLBP. On the other hand, studies claiming acupuncture to be an effective treatment all had a number of severe limitations (see Table 2). Because of this, we must be cautious when using the results from these studies in clinical applications.
We determined that failure to establish (or state) consistent acupuncture treatment protocols was the limitation that would affect the results the most. Acupuncture is traditionally performed as a response to the body’s needs at the time of treatment, which may or may not be the same for all subjects. However, in a controlled study, it is crucial to eliminate such blatant confounding variables as treating the subjects differently. Of the studies reviewed, only Coan et al. (1980) failed to specify a treatment protocol. Some subjects (N was not specified) received both traditional meridian acupuncture and electroacupuncture (a different, more contemporary form of acupuncture), whereas other subjects had only one or the other. In addition, the frequency and duration of treatment sessions for acupuncture patients were not stated. We cannot assume that all subjects were treated equally without a sufficient description of their methods. This study stated that acupuncture was an effective treatment for cLBP (see Figure 3). However we believe that the results of the study have very little internal validity and speculate that had treatment protocols been better controlled, the significant effects of acupuncture may not have existed.
The second most influential limitation was determined to be the consistency of the same acupuncturist for each subject. There is unavoidable therapist-therapist variability, which must be controlled to produce strong data. Therapists will likely treat somewhat differently even if presented with the same symptoms. Similar to not stating treatment protocols, we believe that using different acupuncturists on different subjects implies differing treatment among subjects, which is an obvious confounding variable. Meng et al. (2003) found acupuncture to be effective and had a well-designed treatment protocol (see Figure 4). However, they used more than one acupuncturist to treat. Their protocol consisted of a rigid set of mandatory points (that were applied to all subjects regardless of symptoms at the time of treatment) with up to 4 additional points left up to the discretion of the therapist. Had only one acupuncturist been used this protocol and the results they found would have been much stronger. Therefore, it is difficult to apply the results of this study to the greater public due to our uncertainties of the reliability of the results.
The third most important limitation we found was the extent to which a study’s design was blinded. Blinding has been historically difficult in studies examining the effectiveness of acupuncture for treating cLBP. Typically, the strongest blinding design possible is double-blinded, where both the researcher recording pain levels and the patient are unaware of whether the treatment is traditional or sham acupuncture. Many of the studies included in this review claimed to have double or single-blinded trials. However, none of the studies making such claims described their blinding techniques. We suggest that the problem revolves around the lack of a clear definition of blinding. Because ‘double-blinding’ is ambiguous throughout the literature, we decided that failure to double blind is not a limitation of an individual study, but rather an inherent flaw in the body of knowledge. Two studies, however, (Gunn et al., 1980, and Coan et al., 1980) failed to blind their subjects at all. Both studies reported acupuncture to be an effective treatment of cLBP (see Figure 3 and Figure 5). It is impossible to rule out the chance of a placebo effect in the subjects of these studies because they knew the treatment group they were assigned to. Any prior exposure to acupuncture or any opinions the subjects may have on its effectiveness likely affected the results, which makes application of these results unreliable.
Another limiting factor in several of the reviewed studies was the use of small treatment group sizes. According to Smith et al. (1999), trials with a minimum group size of 40 subjects are considered reliable. Although a small treatment group size is traditionally a large limiting factor in determining statistical significance of outcome measures, the majority of the reviewed studies did not meet this criteria. Because of the lack of reliable data available, we did not weigh the studies with treatment groups of less than 40 subjects as a large limitation. There were only a few studies that actually achieved groups with over 40 participants. These include Mendelson et al. (1983), Molsberger et al. (2002) and Leibing et al. (2002). Mendelson et al. (1983) and Leibing et al. (2002) both concluded that acupuncture was not effective (see Figure 6 and Table 3) and Molsberger et al. (2002) speculated that acupuncture would be an effective treatment (see Table 4). A summary of the treatment group sizes of the reviewed studies can be found in Table 5. It is important to note that many of the studies made strong conclusions in favor of or against acupuncture as an effective mode of treatment of cLBP, when in reality, with such small treatment group sizes, researchers should only make comments regarding trends in the data. Perhaps so many of the current studies lack sufficient participants because acupuncture is time-consuming with an average session lasting approximately an hour. Additionally, several sessions are traditionally used in treatment. Future studies should use treatment group sizes of over 40 subjects in order to verify any statistical significance.
Despite the above limitations, some studies have succeeded in creating strong study designs. The Molsberger et al. study (2002), (which found no significant changes with acupuncture treatments although positive trends were noted) was very well designed as it was randomized, blinded, controlled, and had planned for a long-term study with follow-up visits. Unfortunately, due to external reasons, the hospital where the study was held closed after 1.5 years of research. The researchers then performed extensive analysis on the results to examine all possible angles of the data had the study run its full length, speculating that the same trends would have existed. However, application of knowledge cannot be built solely on assumptions and speculations and although the trends of the study are intriguing, they are not powerful enough to warrant clinical recommendations.
Clinical designs are highly variable concerning control groups across the literature. Many researchers utilize sham acupuncture to aid blinding, which is the placement of needles on inactive points rather than along the traditional Chinese meridians (although other methods have been approached, such as the “fake needle.”). However, none of the reviewed studies utilized a “fake needle” method. Within the current body of knowledge, there are four different design approaches concerning control groups: using sham acupuncture with no control, no sham acupuncture and no control, a control with no sham acupuncture, or a control and sham acupuncture. The strongest studies used both control and sham acupuncture groups, because they are the only ones capable of making implications about potential placebo effects. The results from those that used neither a control group or a group receiving sham acupuncture can only be comparative across the different treatments. A summary of the design approaches concerning control groups can be found in Table 6. Two studies employing the best design (Giles and Muller, 1999, and Leibing et al., 2002) found that acupuncture is not effective in terms of treating cLBP (see Table 3 and Table 7). Conversely, Molsberger et al. (2002) also used both a control and sham acupuncture group, but speculated acupuncture to be effective. However, as their study was not completed, the results do not carry the same weight as the results from Giles and Muller (1999) and Leibing et al. (2002). The use of both control and sham groups is essential in studies focused on acupuncture because the placebo effect can be accounted for. The primary speculation among patients receiving acupuncture treatment in Western cultures relies on this notion. However, the majority of the reviewed studies do not utilize this technique, which reinforces the fact that the current knowledge is limited.