Acupuncture is a traditional Chinese medicine technique that involves procedures to stimulate points on the body. Acupuncture frequently involves penetrating the skin with metallic acupuncture needles (NCCIH, 2018d). Acupuncture has been examined across a wide range of pain conditions from treatment of acute pain in the emergency room to chronic low back pain and cancer-related pain.

Meta-analysis of 14 RCTs supported acupuncture for several acute pain conditions treated in the emergency department. Acupuncture reduced fracture pain the most with a large standard mean difference (SMD) of 2.06 (confidence interval [CI] = 1.43-2.69), had a medium effect on acute back pain with an SMD of 0.75 (CI = 0.03 to -1.48), a medium effect for migraine pain with an SMD of 0.60 (CI = 0.18-1.03), and a small effect for renal colic pain with an SMD of -0.21 (CI = -0.86 to 0.43). Acupuncture reduced acute pain to a level similar to analgesics and when coadministered with analgesics produced an additive effect. Ear acupuncture had a greater effect on pain reduction than acupuncture at other sites; however, no direct comparisons were made. Research that used the same acupoints without needle penetration as the comparison group had a medium, but smaller effect than research without an acupoint comparison group, perhaps because of more rigorous control of response bias. Adverse effects from acupuncture were relatively low at 5.04%. Heterogeneity (high variability) among the studies included in the meta-analysis remains a concern, however (Jan et al., 2017).

Acupuncture for the treatment of acute surgical pain after back surgery has been supported in a subanalysis of two RCTs within a five RCT meta-analysis. Both studies were from the same investigators. Results supported a moderate effect size of acupuncture on reduced pain intensity with an SMD of -0.67 (CI = -1.04 to -0.31) with no heterogeneity concerns (Cho et al., 2015), providing clearer evidence that acupuncture reduces acute surgical pain.

The effect of acupuncture in treating osteoarthritis pain in adults was examined in a meta-analysis of 12 RCTs. Only three of the RCTs had low risk of bias. Combination of the three trials with a total of 410 adults supported a mean pain intensity reduction of -0.59 (CI = -1.18 to -0.0). Comparison of length of intervention across six RCTs dichotomized as less than 4 weeks and 4 weeks and greater supported that longer duration interventions resulted in a greater reduction in pain intensity, SMD -0.38 (CI = -0.69 to -0.06). Functional mobility was also significantly increased as a result of the acupuncture, along with health-related quality of life (Manyanga et al., 2014).

Treatment of nonspecific chronic low back pain with acupuncture is cautiously supported by meta-analysis results. Subanalysis of 5 studies from a 25-study meta-analysis for the effect of acupuncture on nonspecific chronic low back pain supported that acupuncture compared to no treatment produced a moderate effect
of SMD of -0.72 (CI = -0.94 to -0.49) immediately posttreatment. Furthermore, across four studies, acupuncture compared to sham treatment resulted in a mean 100 mm visual analog scale score reduction of -16.76 (CI = -33.33 to -0.19) (Lam, Galvin, & Curry, 2013). High heterogeneity among the studies and small pooled sample sizes demand cautious interpretation, however.

Acupuncture provides moderate pain reduction for malignancy-related pain, with an effect size of -0.71 (CI = -0.94 to -0.048) as reported in a meta-analysis of 29 RCTs of acupuncture and cancer-related pain. Acupuncture more modestly reduced acute surgical pain after cancer surgery with an effect of -0.40 (CI = -0.69 to -0.10). Results must be interpreted within the context of risk for bias across all studies, unclear description of acupuncture procedures, and wide variability in outcomes (Chiu, Hsieh, & Tsai, 2016). Acupuncture as treatment for pain associated with aromatase inhibitor-induced arthralgia was supported in studies using the Western Ontario and McMaster Universities Arthritis Index (WOMAC) but not the Brief Pain Inventory. Reduction in pain intensity occurred after 6 to 8 weeks of acupuncture treatment (Chen et al., 2017).

Acupuncture for pain reduction from dysmenorrhea was examined in a meta-analysis that included 49 RCTs. As with previous meta-analyses of acupuncture for pain treatment, heterogeneity was a problem. Subanalysis of five studies using the visual analog scale supported that manual acupuncture resulted in a significant reduction in pain intensity, SMD = -1.22 (CI = -1.53 to -0.91), after 180 minutes in a sample of 210 women (Woo et al., 2018).

Laser acupuncture involves laser rather than manual stimulation with needles. For musculoskeletal pain, meta-analysis of 33 RCTs supported a small reduction in pain intensity of SMD = -0.43 (CI = -0.74 to -0.12) at short-term follow-up after laser acupuncture, and a moderate effect of SMD = -0.61 (CI = -1.12 to -0.10) at long-term follow-up when laser acupuncture was compared to placebo conditions. No significant pain reduction was supported in RCTs comparing laser acupuncture to control conditions. High heterogeneity among the included studies and dose and time to effect are all areas that require further examination to guide use of laser acupuncture as an effective pain treatment (Law, McDonough, Bleakley, Baxter, & Tumilty, 2015).

Massage involves the use of hands to systematically manipulate soft tissue with the aim of producing positive effects that can include pain reduction (Crawford et al., 2016). Massage to reduce pain has been studied in acute and chronic pain conditions, and more specifically in postsurgical pain and in cancer-related pain. Meta-analysis of 32 randomized controlled clinical trials testing massage in people with acute or chronic pain supported weak positive evidence that massage reduces pain when compared to a sham or active comparator conditions (e.g., relaxation). When compared to no treatment, massage has a stronger effect in reducing pain. Heterogeneity was high, indicating considerable variability between the combined RCTs, however (Crawford et al., 2016).

Cancer-related pain represents a wide variety of affected body systems and differs further if metastases are present. Thus, heterogeneity remains a common
problem when comparing across RCTs. Meta-analysis of six RCTs with a total n = 370 supported weak positive evidence from massage compared to an active comparator. Reduction of pain intensity was not clinically significant, and heterogeneity among the RCTs was high (Boyd et al., 2016a), however. A larger meta-analysis with 12 RCTs and n = 559 people with cancer-related pain provided sensitivity analyses to further inform clinical practice. Analysis of all 12 studies as well as analysis of the 9 high-quality RCTs supported massage as significantly reducing cancer-related pain. In three RCTs of metastatic cancer-related pain, heterogeneity was acceptable and a significant effect was supported for massage. Breast cancer-related pain was analyzed across four studies and supported massage as significantly reducing pain. Analysis of three studies supported massage as effective 2 weeks after the massage treatment. Pooled results from seven studies supported body massage as effective, pooled results from four studies supported foot reflexology as effective, and pooled results from two studies supported aroma massage as effective in reducing cancer-related pain. Massage dose differed across the RCTs with time M = 29.5 minutes (range = 10-50 minutes), sessions M = 4.5 (range = 1-12 sessions), and duration M = 23.6 days (range = 1-140 days) (Lee, Kim, Yeo, Kim, & Lim, 2015).

Massage as a treatment for acute postoperative pain was examined in two meta-analyses. Meta-analysis of critical or acutely ill adults following thoracic surgery that compared massage to sham or attention control or standard care provided weak evidence for pain reduction when all groups also received analgesics. All but one of the 12 studies exclusively examined coronary artery bypass grafts or valve replacement patients, and the majority of patients were male. High heterogeneity and risk of bias suggest cautious interpretation of the results (Boiter, Gelinas, Richard-Lalonde, & Thombs, 2017). Pain reduction was again supported in a second meta-analysis with seven RCTs (n = 1101) across multiple surgical procedures. Results approached clinical significance, with a visual analog scale score reduction of -19.85 (Boyd et al., 2016b).

Tai chi involves slow and gentle body movements while breathing deeply and meditating on the movement. Different tai chi styles exist such as Yang, Sun and Wu (NCCIH, 2018e). Results from two meta-analyses support tai chi as effective in reducing osteoarthritis pain (Hall et al., 2017; Kong et al., 2016); however, the meta-analyses overlapped and included seven of the same research studies. A moderate short-term effect on osteoarthritis pain reduction occurred across 8 RCTs (SMD = -0.54, CI = -0.77 to -0.30 [Kong et al., 2016]) and 11 RCTs (SMD = -0.66, CI = -0.85 to -0.48 [Hall et al., 2017]). Tai chi significantly reduced pain compared to wait-list control and attention control groups, but was not significantly different when compared to active therapy.