Efficacy of Massage Therapy on Pain and Dysfunction in Patients with Neck Pain: A Systematic Review and Meta-Analysis

Objective. To systematically evaluate the evidence of whether massage therapy (MT) is effective for neck pain. Methods. Randomized controlled trials (RCTs) were identified through searches of 5 English and Chinese databases (to December 2012). The search terms included neck pain, neck disorders, cervical vertebrae, massage, manual therapy, Tuina, and random. In addition, we performed hand searches at the library of Nanjing University of Traditional Chinese Medicine. Two reviewers independently abstracted data and assessed the methodological quality of RCTs by PEDro scale. And the meta-analyses of improvements on pain and neck-related function were conducted. Results. Fifteen RCTs met inclusion criteria. The meta-analysis showed that MT experienced better immediate effects on pain relief compared with inactive therapies (n = 153; standardised mean difference (SMD), 1.30; 95% confidence interval (CI), 0.09 to 2.50; P = 0.03) and traditional Chinese medicine (n = 125; SMD, 0.73; 95% CI 0.13 to 1.33; P = 0.02). There was no valid evidence of MT on improving dysfunction. With regard to follow-up effects, there was not enough evidence of MT for neck pain. Conclusions. This systematic review found moderate evidence of MT on improving pain in patients with neck pain compared with inactive therapies and limited evidence compared with traditional Chinese medicine. There were no valid lines of evidence of MT on improving dysfunction. High quality RCTs are urgently needed to confirm these results and continue to compare MT with other active therapies for neck pain.


Introduction
Neck pain is a very common condition. It has one-month prevalence between 15.4% and 45.3% and 12-month prevalence between 12.1% and 71.5% in adults [1]. Despite its high prevalence, neck pain frequently becomes chronic and affects 10% of males and 17% of females [2].Consequently, neck pain has been a source of disability and may require substantial health care resources and treatments [3][4][5][6].
Massage therapy (MT), as one of the earliest and most primitive tools for pain, has been widely used for neck pain. It is defined as a therapeutic manipulation using the hands or a mechanical device, in which numerous specific and general techniques are used in sequence, such as effleurage, petrissage, and percussion [7]. There are, however, inconsistent conclusions on effects of MT for neck pain. Some prior reviews maintained that there was inconclusive evidence on effects of MT for neck pain [8][9][10][11], but the others suggested that MT had immediate effects for neck pain [12,13]. In addition, most reviews did not include Chinese randomized controlled trials (RCTs) of MT for neck pain due to language barrier or limited retrieving resources [8,9,11,12]. But Chinese MT, as one of the primitive complementary and alternative treatments, has been employed by most Chinese patients with neck pain, and a mass of studies have been reported [10]. They are important for evaluating the evidence of MT for neck pain.
Therefore, we performed an updated systematic review of all currently available both English and Chinese publications and conducted quantitative meta-analyses of MT on neck pain and its associated dysfunction to determine whether MT is a viable complementary and alternative treatment for neck pain.

Materials and Methods
The following electronic databases were searched from their inception to December 2012: PubMed, EMBASE, Cochrane 2 Evidence-Based Complementary and Alternative Medicine Library, China Knowledge Resource Integrated Database (CNKI), and Wan Fang Data. The main search terms were neck pain, neck disorders, cervical vertebrae, massage, manual therapy, Tuina, and random. And we performed hand searches at the library of Nanjing University of Traditional Chinese Medicine. Reference lists of retrieved articles were also screened. No restrictions on publication status were imposed.

Eligibility Criteria.
Only the studies that met the following criteria were included: (1) RCTs of MT for neck pain; (2) neck pain was not caused by fractures, tumors, infections, rheumatoid arthritis, and so forth; (3) MT was viewed as an independent therapeutic intervention for neck pain, which did not combine with other manual therapies such as spinal manipulation, mobilization, and chiropractic; (4) the control interventions included inactive and active therapies; the inactive therapy controls included sham, placebo, no treatment, standard care, and others (i.e., massage + exercise versus exercise); the active therapy controls may be any active treatment not related to MT; (5) the main outcome measures were pain and neck-related dysfunction; no restrictions were set on the measurement tools used to assess these outcomes, since a large variety of outcome measures were employed in the studies; (6) the language was either English or Chinese.

Data Abstraction.
Two reviewers independently extracted data onto predefined criteria in Table 1. We contacted primary authors when relevant information was not reported. Differences were settled by discussion with reference to the original article. For crossover studies, we considered the risk for carryover effects to be prohibitive, so we selected only the first phase of the study. We considered that effects of MT included immediate effects (immediately after treatments: up to one day) and follow-up effects (short-term follow-up: between one day and three months, intermediate-term follow-up: between three months and one year, and long-term follow-up: one year and beyond).

Methodological Quality Assessment. The methodological quality of RCTs was assessed independently in line with
PEDro scale by two reviewers, which is based on the Delphi list and has been reported to have a fair to good reliability for RCTs of the physiotherapy in systematic reviews. And the authors compared the results and discussed difference according to the PEDro operational definitions until agreement was reached. The PEDro score ranged from 0 to 10, and a higher score represents a better methodological quality. A cut point of 6 was used to indicate high quality studies as it has been reported to be sufficient to determine high quality versus low quality in previous studies [14,15]. If additional clarification was necessary, we contacted primary authors.

Data Synthesis and Analysis.
The detailed subgroup meta-analyses were performed based on different control therapies. Each subgroup should include at least 2 RCTs. Standardised mean difference (SMD) was used in metaanalyses because the eligible studies assessed the outcome based on different scales (e.g., VAS 0-10 and VAS 0-100). And the SMD and 95% confidence intervals (CI) were calculated in the meta-analyses. We used the more conservative random effects model to account for the expected heterogeneity. The 2 was used to assess statistical heterogeneity. The reviewers determined that heterogeneity was high when the 2 was above 75% [16]. The Cochrane Collaboration software (Review Manager Version 5.0 for Windows; Copenhagen: The Nordic Cochrane Centre) was used for the metaanalyses.
One study was contacted to request for mean and standard deviation data on primary outcomes [24]. Another trial was contacted to provide details on therapeutic technique and study design [31].

Study Characteristics.
Fifteen eligible studies including 1062 subjects with mean age of 41.9 ± 12.4 were, respectively, conducted in Australia, China, Finland, Germany, Poland, Spain, USA, and UK between 2001 and 2012. The disease duration ranged from 1 week to 11.2 years and the study duration 1 day to 10 weeks. The session and time of MT, respectively, were 8.1 ± 5.6 (range 1-18) and 31.1 ± 11.7 minutes (range 20-60 minutes). The follow-up time ranged from 6 to 48 weeks.
MT in the studies included Chinese traditional massage, common Western massage, manual pressure release, strain/counterstrain technique, and myofascial band therapy. The control therapies contained inactive therapies (standard care and sham therapies) and active therapies including acupuncture, traction, physical therapy, exercise, traditional bone setting, traditional Chinese medicine, joint mobilization, and activator trigger point therapy. The characteristics of all studies were summarized in Table 1.

The Effects of MT on Dysfunction.
Six RCTs examined the immediate effect of MT on dysfunction by neck disability index (NDI) versus inactive therapies [24,31] or active therapies [21,23,26,27]. All of them were included in the meta-analysis. The aggregated results suggested that MT did not show significant immediate effects on dysfunction 6 Evidence-Based Complementary and Alternative Medicine      Figure 5).
Four RCTs assessed the immediate effect of MT on range of motion of the neck compared with exercise (or standard care) [18], acupuncture [27], traditional Chinese medicine [29], and physical therapy [31].

Adverse Events.
Only two studies reported side effects. One study reported that 21% of the participants experienced low blood pressure following treatment [17]. The other trial reported that 9 (about 28%) participants had mild adverse experiences including discomfort, pain, soreness, and nausea [24].

Discussion
The purpose of our systematic review was to evaluate the evidence of MT for neck pain. Our meta-analyses found beneficial evidences of MT for neck pain. Compared with 8      inactive therapies, MT showed moderate evidence for immediate improvement of pain, and compared with traditional Chinese medicine there was limited evidence for immediate improvement of pain due to few eligible studies. However, MT did not show better effects versus other active therapies (including acupuncture, traction, and other manual therapies). And there was no evidence that MT showed superior immediate effects on improving dysfunction in patients with neck pain. On follow-up effects, there was not enough evidence of MT for neck pain. Our review contained six Chinese RCTs of MT for neck pain. Although MT is widely used for neck pain in China, most of the previous reviews included few Chinese RCTs of MT for neck pain due to limitations of retrieving resources and methodological qualities. In our review, all Chinese RCTs performed eligible random allocation and the quality scores were more than 6 in terms of PEDro scores. They failed to blind the subjects and therapists, but three RCTs [27][28][29] performed eligible concealed allocation, and one [23] employed blinded assessors. What is more, it is difficult to blind the patients and therapists in MT studies. In general, methodological quality of Chinese RCTs of MT for neck is becoming better.
In our review, there were more detailed subgroup analyses based on inventions of control groups. In order to address the question of what her MT is an effective therapy for neck pain, we analyzed studies comparing MT with inactive therapies including sham therapies and standard care. The result only showed that MT may be more effective than standard care. And we also compared MT with active therapies including acupuncture, traction, traditional Chinese medicine, physical therapy, exercise, and other manual therapies for assessing the question of what her MT is a better therapy for neck pain. The meta-analysis showed that MT has better immediate effects than traditional Chinese medicine, but eligible studies were few. And the treatment process of traditional Chinese medicine is usually longer; 3 to 4 weeks of traditional Chinese medicine may be shorter for neck pain [25,29]. So we considered that MT did not show better effects than other active therapy. In addition, we also paid attention to dysfunction related neck pain and follow-up effects of MT for neck pain. Heterogeneity: 2 = 0.12; 2 = 7.18, df = 3 (P = 0.07); I 2 = 58% Test for overall effect: Z = 1.01 (P = 0.31) Heterogeneity: 2 = 0.09; 2 = 6.12, df = 3 (P = 0.11); I 2 = 51% Test for overall effect: Z = 1.45 (P = 0.15) Heterogeneity: 2 = 0.01; 2 = 3.28, df = 3 (P = 0.35); I 2 = 9% Test for overall effect: Z = 1.84 (P = 0.07) Heterogeneity: 2 = 0.00; 2 = 2.43, df = 3 (P = 0.49); I 2 = 0% Test for overall effect: Z = 0.91 (P = 0.36) Heterogeneity: 2 = 0.08; 2 = 30.26, df = 15 (P = 0.01); I 2 = 50% Test for overall effect: Z = 0.66 (P = 0.51) Test for subgroup differences: 2 = 5.40, df = 3 (P = 0.14); I 2 = 44.5% −0. 18   And it supported the effectiveness of massage for neck pain remained uncertain [8]. Its result concurred with the result of our review, but our review excluded a few studies that Patel had included because they used treatments related to MT in control groups [55][56][57][58]. These were limited to evaluating the specific effect of MT. And some studies were not eligible for inclusion criteria of our review [59][60][61][62]. Moreover, our systematic review included eight new RCTs [23,[25][26][27][28][29][30][31] published from 2008 to 2012. Of notes, our review contained six Chinese RCTs of MT for neck pain [23,25,[27][28][29][30]. And we assessed the effect of MT on neck pain and its associated dysfunction. We also paid attention to the immediate and follow-up effects of MT. So our update provides stronger evidence of MT for neck pain. Our results differ from systematic reviews [12,13]. Ottawa panel evidence-based clinical practice guidelines, including five RCTs with high methodological quality (>3) according to the Jadad scale, suggested that MT was effective for relieving immediate posttreatment neck pain symptoms [12]. One suspected reason for this difference is that a mass of new RCTs [20,21,23,[25][26][27][28][29][30][31] have been published, which were not included in their review. Another possible explanation for the difference is that Jadad scale was replaced by PEDro scale in our review, which is a more detailed method based on the Delphi list and has been reported to have a fair to good reliability for RCTs of the physiotherapy in systematic reviews. In addition, detailed meta-analyses were performed based on more RCTs in our review. Ottawa panel clinical practice guidelines declined to combine the trials because of fewer trials. Moreover, we separately compared MT with inactive therapies and active therapies, and assessed the effect of MT on neck pain and its associated dysfunction in our review. More eligible RCTs, classification of quantitative data synthesis, and detailed assessment of MT on neck pain and its associated dysfunction strengthened our confidence in our systematic review.

Limitations.
There are several limitations in our review as follows. (a) Although the predetermined cutoff 6 was exceeded, there were serious flaws in blinding methods of most Chinese RCTs. It is difficult to blind the patients and impossible to blind the therapists, but blinded assessors and concealed allocation must attempt to make up for the lack of blinding. However, some Chinese RCTs did not perform these compensated methods. Thus, these studies could not be considered to be of high quality. (b) Our review may also be affected by dosing parameters of MT such as duration (time of each MT), frequency (sessions of MT per week), and dosage (size of strength). MT commonly combines different techniques (stroking, kneading, percussion, etc.), and each therapist may perform them in different dosing parameters. So the dose-finding studies are warranted to establish a minimally effective dose. (c) The results may be influenced by different outcome measures of pain and dysfunction in eligible RCTs. So the reliable and valid outcome measures is essential to reduce bias, provide precise measures and perform valid data synthesis. (d) There were less eligible trials in some subgroups of meta-analyses because of strict eligibility criteria for considering studies in our review. It may influence combining results, but low eligibility criteria would generate more doubtful results. (e) The majority of trials did not report adverse events, so it was not clear from the reports whether adverse effects had been measured or not.

Conclusions
Although there were no valid lines of evidence of MT on improving dysfunction in patients with neck pain, this systematic review found moderate evidence of MT on improving pain in patients with neck pain compared with inactive therapies and limited evidence compared with traditional Chinese medicine due to few eligible studies. These are beneficial evidence of MT for neck pain. Assuming that MT is at least immediately effective and safe, it might be preliminarily recommended as a complementary and alternative treatment for patients with neck pain. But more high quality RCTs are urgently needed to confirm these results and continue to compare MT with other active therapies for neck pain.