Acupuncture for Spinal Cord Injury and Its Complications: A Systematic Review and Meta-Analysis of Randomized Controlled Trials

To evaluate the evidence supporting the effectiveness of acupuncture treatment for SCI and its complications, we conducted search across 19 electronic databases to find all of the randomized controlled trials (RCTs) that used acupuncture as a treatment for SCI and its complications. The methodological quality of each RCT was assessed using the Cochrane risk of bias tool and the PEDro scale. Sixteen RCTs, including 2 high-quality RCTs, met our inclusion criteria (8 for functional recovery from SCI, 6 for bladder dysfunction, and 2 for pain control). The meta-analysis showed positive results for the use of acupuncture combined with conventional treatments for the functional recovery in terms of motor ASIA scores and total FIM scores when compared to conventional treatments alone. Positive results were also obtained for the treatment of bladder dysfunction, in terms of the total efficacy rate, when comparing acupuncture to conventional treatments. However, 2 RCTs for pain control reported conflicting results. Our systematic review found encouraging albeit limited evidence for functional recovery, bladder dysfunction, and pain in SCI. However, to obtain stronger evidence without the drawbacks of trial design and the quality of studies, we recommend sham-controlled RCTs or comparative effectiveness research for each condition to test the effectiveness of acupuncture.


Introduction
Spinal cord injury (SCI) affects approximately 900 to 1,000 individuals per million in the general population, and it is estimated that there are 12,000 new cases of SCI every year in the United States [1]. Nearly 80% of the individuals who experience SCI are male, and, since 2000, the average age at injury has increased to 39.5 years, with 11.5% of those injured greater than 60 years of age [2]. These statistics indicate that SCI may be a significant social and economic burden for patients and their families.
SCI causes lesions, damaged neurological structures, and secondary pathophysiological changes in spared tissue [19]. Thus, complete or partial loss of sensory and motor function is the most significant result of injury, and most patients experience secondary complications, such as bladder and bowel dysfunction, chronic pain, infertility, and autonomic dysfunction [19].
There are many treatments for functional recovery from injury and the related complications that result from SCI; these include surgery [20], prescription drugs [21], behavioral therapy [22], physical therapy [23], and supportive treatment [24]. There are also a variety of treatments used for the secondary complications of SCI, such as intermittent catheterization for bladder dysfunction, analgesics for pain, and others [25]. These treatments tend to be administered over long periods of time, and because of the potential complications of treatment, there has been an increased interest in alternative medical treatments, including acupuncture and other related therapies (moxibustion and acupressure) [26]. Acupuncture 2 Evidence-Based Complementary and Alternative Medicine based on Traditional Chinese Medicine has been commonly used for pain or neurological problems in Chinese cultures [27]. Additionally, many studies [28,29] have analyzed the use of acupuncture for these types of problems; these studies report a variety of outcomes regarding level of function, pain, and quality of life.
Three recently published reviews have examined the results of studies that support the use of treatments such as acupuncture for individuals with SCI [26,30,31]. One review [30] that was limited to a Chinese literature suggested that acupuncture treatments have a positive effect, while a second [26] did not systematically evaluate the available evidence and even failed to include several important trials [3, 5-13, 15, 16, 18]. The third review was restricted to the pain related to SCI [31]. Therefore, the purpose of this systematic review was to provide a qualitative analysis of all of the randomized controlled trials (RCTs) to date that were designed to determine the effectiveness of acupuncture for patients with SCI and its complications along PRISMA guidelines [32].

Data Sources.
This systematic review included studies published in electronic databases over the time period ranging from their inception to December 2011. Relevant publications were found in the following databases: the Cochrane Central Register of Controlled Trials (CENTRAL), the Cochrane Database of Systematic Review (CDSR), PubMed, MEDLINE, EMBASE, CINAHL, nine Korean databases (Korean Studies Information, DBPIA, Korea Institute of Science Technology Information, Korean National Assembly Library, RISS4U, KoreaMed, Korean medical database, Korean Traditional Knowledge Portal, and Oasis), Chinese database (China National Knowledge Infrastructure), and two Japanese databases (J-STAGE and JAMAS). In addition, we searched databases that contained registered trials, such as ClinicalTrials.gov (http://www.clinicaltrials.gov).
The search keywords used were (acupuncture OR acup * OR "electroacupuncture" OR "auricular acupuncture" OR "scalp acupuncture") AND ("spinal cord injury" OR "spinal injury" OR "spinal cord trauma" OR "spinal cord lesion" OR "spinal cord damage" OR "spinal cord fracture" OR "spinal cord contusion") in each base language. This search strategy was adjusted for each database. In addition, the bibliographies of relevant systematic reviews and clinical guidelines were manually searched. We also searched the gray literature that included dissertations, letters, government documents, research reports, conference proceedings, and abstracts when available. The reference section for each study was also searched.

Types of Studies.
We evaluated RCTs that studied the clinical effects of acupuncture as a treatment for SCI or its direct complications. Dissertations and abstracts examining these topics were also included. The review included prospective, parallel RCTs, or cross-over RCTs that assessed the efficacy of acupuncture regardless of blinding, language, or the type of report. The title and abstract of each identified article were read by a single primary researcher (IH) who completed the screening process. If articles were not written in English, they were translated into English prior to screening. The articles that would then be potentially included in our analysis were carefully checked by 2 independent reviewers (IH, YDK).

Types of Participants.
The clinical trials included in our review examined patients with SCI or complications secondary to SCI. There were no restrictions related to the amount of time after injury, the type of injury, the site of injury, or participant's age.

Types of Interventions.
Our study considered the effects of needle acupuncture. We included needle acupuncture with or without electrical stimulation or heating by moxa, auricular acupuncture, and scalp acupuncture. We excluded injection acupuncture, nonneedling acupuncture (e.g., laser acupuncture), acupuncture-like intervention (e.g., acupressure, moxibustion), and mixed treatments. Control conditions in the reviewed studies included sham acupuncture, no treatment, and conventional treatment for SCI (e.g., medication, rehabilitation). Studies using cointerventions were included only if cointerventions were given to both treatment groups.

Outcome Measures.
The outcome measures we considered were neurologic status or score (i.e., the American Spinal Injury Association neurologic and functional score; the ASIA score), functional ability outcomes (i.e., the Fugl-Meyer score), activities of daily living (i.e., the Functional Independence Measure; FIM score), scores related to the efficacy rate (i.e., the rate of participants who demonstrated efficacy), and outcome measures related to the complications of SCI (i.e., the bladder function scale for bladder dysfunction, pain scores for the level of pain and range of motion; ROM). We also considered measures of general safety reported for acupuncture as a treatment.

Data Extraction and Quality Assessment.
Two reviewers (IH, YDK) reviewed each article independently and were blinded to the findings of the other reviewer. The reviewers collected data relating to the methodology, outcome measures, results, and final conclusions of each study. The reviewers also quantitatively evaluated each study's methodological quality through the use of the Cochrane risk of bias [33] and the Physiotherapy Evidence Database (PEDro) scale [34], which was developed to assess the methodological quality of RCTs specifically pertaining to physical therapy interventions [35]. The PEDro scale allows researchers to assess the quality of studies based on a cutoff score: studies that score fewer than six points are considered to be "low quality, " while studies with a score equal to or greater than six points are considered to be "high quality" [36]. Any discrepancies Evidence-Based Complementary and Alternative Medicine 3 between reviewers were resolved through discussion until a consensus was reached.

Data Synthesis.
To summarize the effects of acupuncture on each outcome measure, we used Cochrane Collaboration software (Review Manager (RevMan) Version 5.1 for Windows. Copenhagen: The Nordic Cochrane Centre). We extrapolated the risk estimate (relative risk; RR) and the 95% confidence interval (CI) for dichotomous data. The standard mean difference (SMD) and the 95% CI were calculated for continuous data. The variance of the amount of change was imputed using a correlation factor of 0.4, which is the value suggested by the Cochrane Collaboration [33]. We pooled data across studies using random effect models if excessive statistical heterogeneity did not exist. The chi-squared test and the Higgins 2 test were used to assess the heterogeneity of the data [33]. We planned to use a formal funnel plot to assess publication bias if more than 10 trials were included; however, we were unable to generate a funnel plot due to the small number of trials included in our meta-analysis.

Study
Quality. The mean PEDro score was 4.5; scores ranged from 4 to 8 points (Table 2). Only 2 of 16 RCTs met the PEDro criteria for high quality [9,17]. The results of the Cochrane risk of bias analysis varied widely ( Table 2). Two studies reported appropriate sequence generation [17,18], in which the researchers used a coin toss method [18] and stratified block randomization [17]. Two of the trials reported inappropriate randomization procedures [8,16], in which the researchers used hospital admission numbers for randomization. One trial blinded the participants and the assessors [17], and three blinded the assessors only [4,9,18]. Only one trial mentioned that the outcomes were analyzed with an intention-to-treat analysis [17]. Incomplete outcome data were addressed adequately in three studies [14,17,18]. Only one trial implemented allocation concealment; however, a detailed description of the method was not reported [9].

Functional Recovery
Acupuncture versus Conventional Treatment (1 RCT). One RCT evaluated the effect of electrical acupuncture compared to rehabilitation [3]. A significant difference was found between the two groups regarding the total efficacy rate ( = 0.05, Table 1). (7 RCTs). Five RCTs evaluated the effects of electrical acupuncture plus rehabilitation and compared the results to those of rehabilitation alone [4][5][6][7]9]. Three RCTs described improved outcomes with electrical acupuncture plus rehabilitation over rehabilitation alone [4,7,9]. However, 2 RCTs did not show any significant differences between treatment groups (Table 1) [5,6].

Acupuncture Plus Conventional Treatment versus Conventional Treatment
One RCT compared the effects of electrical acupuncture, rehabilitation, and oral neurotropic drugs to those of rehabilitation and oral neurotropic drugs [8]. This study showed a positive effect in terms of rehabilitation effectiveness only ( = 0.05, Table 1).
One RCT evaluated the effects of electrical acupuncture plus intravenous drugs compared to intravenous drugs alone [10]. This RCT showed that treatment consisting of electrical acupuncture plus intravenous drugs was more effective than intravenous drugs alone ( = 0.003).
The meta-analysis of acupuncture plus conventional treatment versus conventional treatment alone showed that treatment including acupuncture led to significantly more improved motor ASIA scores (2 studies [4,7]: = 156, SMD = 0.78, 95% CI of 0.36 to 1.20, = 0.0002, heterogeneity:     The author did not report total FIM score but did each of 6 domains. † 80 patients were randomized, but 60 were analyzed. ‡ Each group had a significant effect after therapy ( < 0.01).

8
Evidence-Based Complementary and Alternative Medicine
Two of the 3 RCTs used intramuscular neostigmine methylsulfate plus intermittent catheterization and bladder training [11,12]. One study showed that electrical acupuncture induced significantly greater improvement when compared to conventional treatment alone in terms of the total efficacy rate ( = 0.02, Table 1) [11]. However, the other study did not show a significant difference between treatment conditions ( = 0.11, Table 1) [12].
One of the 3 RCTs compared the effects of electrical acupuncture with conventional treatment (intermittent catheterization and bladder training) [13]. This study showed a significant positive effect for electrical acupuncture in terms of the total efficacy rate ( = 0.006, Table 1) and the levels of residual urine ( < 0.00001, Table 1) [13].
The meta-analysis of acupuncture versus conventional treatment showed a significant positive effect of acupuncture in terms of the total efficacy rate (3 studies [11][12][13]: = 264, RR= 1.32, 95% CI of 1.06 to 1.65, = 0.01, heterogeneity: 2 = 4.70, = 0.10, and 2 = 57%; Figure 2). (3 RCTs). Two RCTs compared the effects of electrical acupuncture plus intermittent catheterization and bladder training to intermittent catheterization and bladder training alone [14,16]. A recalculation of the mean difference (MD) revealed that there was a significant positive effect for electrical acupuncture in terms of the total days needed to reach bladder balance ( < 0.05, Table 1) [14]; however, there was not a positive effect in bladder voiding function parameters (the frequency of urination, the maximum voided volume, bladder capacity, residual urine, and quality of life scores, Table 1) [16].

Acupuncture plus Conventional Treatment versus Conventional Treatment Alone
One RCT compared the effects of electrical acupuncture plus intermittent catheterization versus intermittent catheterization alone [15]. This study showed a positive effect in terms of the total efficacy rate ( = 0.007, Table 1) [15].

Pain Control
Acupuncture versus Sham Acupuncture (1 RCT). One RCT evaluated the effects of manual acupuncture with no manipulation at sites located at least 1 cun (1 Chinese anatomic inch, approximately 2.5 cm) away from the established meridian and extra points [17]. The results showed that acupuncture did not significantly affect the PU-WUSPI and the NRS (Table 1) [17]. (1 RCT). One RCT assessed the effects of manual acupuncture compared to the Trager Approach (a form of bodywork and movement reeducation developed by Milton Trager) [18]. Our recalculation of the MD showed that there was no significant difference between the 2 treatments ( > 0.05, Table 1) [18].

The Safety
Reporting of Acupuncture. Only three RCTs reported adverse events associated with acupuncture [7,17,18]. Two RCTs reported no adverse events [7,18], and 1 trial reported minimal adverse effects without a detailed explanation [17].
Based on the Cochrane risk of bias [33] and the PEDro scale [34], the methodological quality and design of the primary studies was mostly poor (only 2 of 16 were considered high quality, 12.5%). Of the 16 studies included in our review, 2 trials used appropriate sequence generation methods [17,18]. Only one of the studies reported allocation concealment [9]. One RCT reported details of assessor-and patientblinding procedures [17]. Trials with inappropriate randomization were threatened by selection bias [50] and inadequate blinding tended to exaggerate the effects of treatment [33]. Researchers should conduct their trials according to the CONSORT statement [51]. In addition, 14 of the 16 RCTs originated from Chinese sources [3][4][5][6][7][8][9][10][11][12][13][14][15][16]. Several groups have shown that the majority of Chinese acupuncture studies report positive results [52,53]. Therefore, it is possible that a publication bias exists; although we searched extensively for all the studies that are relevant to this paper, we may have failed to conduct an analysis of publication bias [54]. This phenomenon casts a considerable doubt on the reliability of these data.
Ten trials compared acupuncture plus various conventional therapies to conventional therapies alone [4][5][6][7][8][9][10][14][15][16], and 5 trials compared acupuncture to conventional therapies [3,[11][12][13]18]. Such trial designs are open to bias because participants were not blinded. Sham-controlled trials could control nonspecific effects of acupuncture [55]. Only one trial adopted a penetrating sham control [17]. This shamcontrolled trial had a small sample size, and our recalculated power for this study was 0.326. This finding indicates that the study lacked sufficient statistical power. Thus, these results limited our ability to evaluate the effectiveness of this treatment for various conditions in SCI. Additionally, nonpenetrating sham acupuncture may be more acceptable than penetrating sham [56,57].
Although the effectiveness of acupuncture is unclear in SCI or its complications, acupuncture therapy is a relatively convenient and safe treatment for some conditions [58]. SCI patients who receive long-term rehabilitation or medication treatments may occasionally need a safe nonpharmacological treatment [59]. In this regard, acupuncture can be useful to treat SCI and its complications if patients experience side effects or have no (or a weak) response to a conventional treatment. Thus, we carefully present several points that must be discussed in future research.
First, although several animal studies of SCI have reported that acupuncture induces neuronal function recovery [60], an analgesic effect [61], and anti-inflammatory responses [62], the RCTs included in this paper might have used specific treatment conditions that may not have fully drawn the maximal benefit from acupuncture. To evaluate the effects of acupuncture on functional recovery in SCI, an appropriate set of treatment conditions which can maximize the therapeutic effect of acupuncture is highly recommended. Also, the conditions have to include the appropriate duration of treatment resulted from comparison of duration of the course of acupuncture treatment. Acupuncture as a (sole or adjunct) treatment for bladder dysfunction demonstrated a positive effect over all conventional treatments except one; while this is encouraging, this result is inconclusive due to the small sample size and low methodological quality. The sham trial for pain control that was reviewed failed to show any specific positive analgesic effect of acupuncture; however, this trial lacked statistical power. It is possible that acupuncture is an effective treatment for pain in SCI, as indicated by a previous rigorous review [31]. Thus, future high-quality trials on secondary conditions of SCI should be conducted to evaluate the potential effects of acupuncture [63].
Second, the prescription of acupuncture points was not consistent across studies (Table 3). Also, the degree of stimulation and the duration and frequency of acupuncture treatment were very various (Table 3). And we could not find the consistency of the Chinese medicine pattern in included trials. So it was difficult to estimate the correlation between the difference of acupuncture treatment and its therapeutic effectiveness. They might not have been optimal or might even be considered an underdosage of treatment for SCI [64]. As a matter of fact, the optimal prescription of acupuncture treatment (acupuncture point, degree of stimulation, frequency of treatment, and a number of treatment sessions) for SCI is a controversial issue amongst acupuncture experts. A standardized prescription of acupuncture for SCI or its complications is necessary.
Third, sham controls in acupuncture research for SCI may represent an ethical issue [65]. Therefore, an adjunctive treatment of acupuncture to conventional treatments for SCI can be evaluated by RCTs comparing acupuncture plus conventional treatment to sham acupuncture plus conventional treatment or through comparative effectiveness research (CER) [66].

Conclusion
The results of our systematic review and meta-analysis suggest that the evidence for the effectiveness of acupuncture as a symptomatic treatment for SCI and its complications is encouraging but limited. There is a great need to test the clinical implications of acupuncture for a number of SCI-related conditions. The efficacy of acupuncture for SCI and its complications must be studied using sham-controlled RCTs or CER with a standardized acupuncture procedure; such RCTs would conform to the recommendations of the CONSORT and STRICTA guidelines [51,67].