This systematic review was registered with the Open Science Framework (10.17605/OSF.IO/NWGSF).¹ PRISMA guidelines were followed (Moher et al., 2009). PubMed, Embase, the Cochrane Library, and EBSCO were searched from January 1991 to March 2022 for relevant clinical trials (Supplementary material 1). The following combination of terms was used as search keywords in the title and abstract: T1 = Tai Chi OR “Tai Chi *” OR Qigong OR Liuzijue OR Wuqinxi OR Yijinjing OR Baduanjin OR “traditional exercise” OR traditional Chinese medicine OR “Chinese traditional exercise” OR “traditional Chinese exercise” OR “Chinese exercise,” T2 = back pain OR low backache OR lower back pain OR lumbago OR lumbosacral pain OR sciatica. When screening clinical trials, the inclusion criteria are as follows:

Two authors independently screened all titles, abstracts, and main text of the relevant papers according to the inclusion criteria. Papers that did not match the criteria for inclusion were omitted. Disagreements were settled by discussion or a third reviewer. The following information was extracted from the selected articles: (1) published data (author, year); (2) design of included studies (subject subgroup, sample size, randomization, follow-up, clinical outcome measures, and time points); (3) type of intervention (including dose regimen, duration); (4) characteristics of participants (including baseline demographic information and diagnostic/inclusion/exclusion criteria); and (5) adverse effects.

We used the Physiotherapy Evidence Database scale (PEDro) to assess the risk of bias for inclusion and the methodological quality of each study in this systematic review (Supplementary Table 1). Two authors independently evaluated the quality of the included RCTs, and all disagreements were settled by discussion or a third reviewer. The following information was evaluated: randomized allocation, concealed allocation, baseline comparability, blind subjects, blind therapists, blind assessors, adequate follow-up, intention-to-treat analysis, between-group comparisons, point estimates, and variability. Scores < 4 points were considered as poor quality; 4–5 points as modern quality; 6–8 points as high quality; 9–10 as excellent quality. The characteristics of the included RCTs were examined (Table 1). Then, The findings were then narratively presented in terms of the TCE’s mechanisms in the treatment of LBP, which were detailed and discussed in the following sections.

As shown in Figure 1A, 718 papers in related fields were retrieved from the four electronic databases. After removing 235 duplicates, 483 articles were screened for eligibility. Through reviewing the titles, abstracts, and full contents of the selected studies, we excluded another 472 articles (review = 62, protocol = 31, animal studies = 14, non-LBP = 93, non-TCE = 179, non-RCT = 80, no essential outcomes = 6, age of subjects less than 35 years old = 7). Finally, 11 RCTs were included in this review (Hall et al., 2011, 2016; Blodt et al., 2015; Teut et al., 2016; Qaseem et al., 2017; Liu et al., 2019; Phattharasupharerk et al., 2019; Zou et al., 2019a; Ma et al., 2020; Sherman et al., 2020; Yao et al., 2020).

Table 1 shows the key characteristics of all included RCTs. Eleven RCTs involving 1,256 participants (810 females) ranging in age from 35 to 73 years were included. The sample size of each RCT ranged from 43 to 176. These RCTs were performed in China, Germany, Australia, England, Thailand, and the USA between 2013 and 2020 (Figure 1B). Three kinds of TCE programs were used to treat CLBP in all intervention groups (7 for Tai Chi, 3 for Qigong, and 1 for Wuqinxi). In the control group, active interventions (such as core training, exercise therapy, or yoga) or passive interventions (such as health education) were used. The treatment duration ranged from 6 weeks to 6 months, with each session lasting from 30 to 90 min, of which 12 weeks was the most common in six trials. The visual analogue scale (VAS) and the numerical rating scale (NRS) were used to evaluated the major outcomes for pain severity.

Supplementary Table 1 summarizes the quality evaluation results for each RCT by using the PEDro scale. The quality of all 11 studies considered in this review ranged from moderate to high (Figure 1C). In most RCTs, participants, therapists, and assessors were not blinded.

All 11 included studies involving 1,256 patients with LBP examined the effect of different types of TCE on pain intensity. Ten studies suggested that TCE group outperformed the control group in terms of pain relief, but one of the studies found that the effectiveness of Qigong for pain relief decreased over time. Only one study found no significant difference in pain alleviation between the Qigong and the control groups (Teut et al., 2016).

Of the 11 included studies, five studies investigated the effects of TCE on back functional disability, three studies evaluated the effect on quality of life (QOL), and four studies determined the effect on sleep quality and satisfaction. Three studies suggested that Tai Chi had effectively improved back pain-related disability compared with the control group (Hall et al., 2011, 2016; Sherman et al., 2020). Phattharasupharerk et al. and Blödt et al. reported that Qigong did not improve back functional disability effectively because the Roland–Morris Disability Questionnaire (RMDQ) scores did not meet the minimal clinically important difference level (Blodt et al., 2015; Phattharasupharerk et al., 2019). Additionally, Qigong and Wuqinxi had effectively improved QOL and depression compared with the baseline, but no statistical difference was found compared with the control group (Blodt et al., 2015; Teut et al., 2016; Yao et al., 2020). Overall, most studies did not find any difference in the secondary outcomes tested (disability, QOL, and sleep quality).

Only two studies reported adverse events. Amanda et al. found that four subjects reported a slight increase in back pain at the beginning of Tai Chi training, which was relieved by the third or fourth week of the training (Hall et al., 2011). Blödt et al. found that both the Qigong group (n = 10) and exercise group (n = 10) reported suspected adverse events (e.g., muscle soreness and tenseness, dizziness, mood fluctuation, and increased back pain) (Blodt et al., 2015).

Pain management in elderly patients with CLBP is particularly challenging, and long-term opioid use was associated with an increased risk of comorbidities (Makris et al., 2014), psychological distress [e.g., depression (Maher et al., 2017)] and other health problems (e.g., falls, osteoporosis, and muscular atrophy) (Chou et al., 2015). A previous review suggested that people who experienced LBP had a higher risk of recurrence (Taylor et al., 2014). Therefore, current guidelines recommend that treatment should focus on reducing pain and its associated dysfunction, and exercise is an effective treatment option (Koes et al., 2010; Zhang et al., 2020). This systematic review included 11 RCTs involving 1,256 patients with CLBP aged over 35 years to assess the overall effect of TCE in middle-aged and elderly patients with CLBP. Our results indicated that TCE could be an effective therapy for reducing pain and improving function in the patients. All treatments (Tai Chi, Qigong, and Wuqinxi) showed positive effects compared with baseline measurements.

Pain intensity relates to the degree to which a person is harmed by CLBP and can be quantified to estimate the severity of pain (Ostelo and de Vet, 2005). The results from this systematic review suggest that TCE significantly reduce the VAS or NRS scores of patients with CLBP. Compared with the control and exercise therapy groups, TCE showed better effects in alleviating pain, which is consistent with prior reviews on other exercise therapies (Hayden et al., 2005; Macedo et al., 2013). Only one study showed that 3 months of yoga or Qigong training had no effect on back pain, back function, or QOL in older patients with CLBP (Teut et al., 2016). However, the number of studies was insufficient to conclude every type of TCE. Additionally, several RCTs showed that Tai Chi/Qigong significantly contributed to proprioception and neuromuscular function in the lower limbs (Zou et al., 2019a). An RCT involving 84 patients showed that Tai Chi was more effective in improving spinal movement function, with shorter training time and better compliance compared with standard exercise therapy. Another RCT showed that office workers with CLBP achieved better health behavior after 6 weeks of Qigong exercise as well as significantly improved mental state, back function, range of motion, and core muscle strength (Phattharasupharerk et al., 2019). Despite the negative findings of some studies, the fact that CLBP is difficult to manage suggests that TCE could be a possible option for managing pain in middle-aged and elderly patients with CLBP.

TCE focuses on the integration of mental regulation, breathing, and movement control in addition to internal energy regulation (Zou et al., 2019b).

First, as a foundation of mind–body interaction, meditation and rhythmic breathing can effectively boost vitality and induce energy to flow through the body, which in turn drives body movement to alleviate pain (Zou et al., 2017b). Self-awareness combined with self-correction of posture and movement of the body, flow of breath, and mental stilling activates natural self-regulatory (self-healing) abilities and stimulates a balanced release of endogenous neurohormones and a variety of natural health recovery mechanisms (Jahnke et al., 2010; Linek et al., 2020). Multiple elements of health, including mood, pain, immunity, and peripheral autonomic nervous system function, can be regulated by concentration and mindful meditation (Wayne and Kaptchuk, 2008). Some TCE programs use meditation and imagination to guide and distract attention away from pain, which can help reduce pain and enhance psychosocial health. Evidence indicates that poor pain-related outcomes (e.g., pain levels and disability) have been linked to a higher level of pain catastrophizing (Peng, 2012). Pain-related catastrophizing is a negative cognitive response to pain. For example, Hall et al. found that Tai Chi could help with pain-related symptoms by changing cognitive appraisal results, such as lowing catastrophic outcomes (Hall et al., 2016). Additionally, a recent meta-analysis suggested that adults with chronic diseases obtained reduced muscle pain through mindfulness-based training (Zou et al., 2018b).

Second, TCE relieves pain by combining muscle strength, static balance, and dynamic balance, and these concepts are quite similar to other pain-relieving therapies, such as core stabilization training (Hall et al., 2011; Gordon and Bloxham, 2016). TCE can help relieve back pain by strengthening lumbar muscles and improving pelvic–lumbar neuromuscular function and proprioception. Qigong, Tai Chi, and Wuqinxi involve a series of slow, flowing, dance-like body movements. In particular, combining slow coordinated postures can transfer upper and lower body momentum and achieve balance depending on continuous squatting and weight shifting on both legs throughout the exercise (Zou et al., 2018b). Improvements in lower limb function and lumbar flexibility improved CLBP-related physical activities (i.e., sitting and standing, stair climbing, and walking) (Masharawi and Nadaf, 2013). Compared with taking anti-osteoporosis drugs, Wuqinxi can significantly reduce pain symptoms and increase the bone density of lumbar vertebrae, suggesting the positive effect of Wuqinxi on CLBP (Wei et al., 2015). Our previous work also suggested that Chen-style Wai Chi had protective effects on neuromuscular function in elderly patients with CLBP while relieving non-specific chronic pain (Zou et al., 2019a).

In addition to correcting postural control and enhancing muscle strength to relieve back pain, the pain-relieving effects of TCE may be linked to changes in TCE-induced brain activity. Long-term Tai Chi practice resulted in an increase in cortical thickness of the inferior segment of the circular sulcus of the insula as well as a decrease in the functional homogeneity of the left anterior cingulate cortex (ACC) (Wei et al., 2013, 2014). ACC plays a crucial role in the emotional aspects of pain (Fuchs et al., 2014; Barthas et al., 2015). Inhibiting ACC may help alleviate chronic pain (Gu et al., 2015). The improvement of ACC functional specificity after Tai Chi training may contribute to pain relief, thereby explaining its analgesic effect. An increase cortical thickness of insula observed in long-term Tai Chi practitioners may also contribute to pain relief through better processing of pain-related cognitive information. To uncover the neurological mechanism underlying TCE-mediated pain relief, further studies should be conducted into the direct relationship between pain perception and TCE-mediated alterations in these brain regions.

This systematic review has some limitations. First, it has location and language bias, with five studies from China, two studies from Germany, and only one study from Thailand, the USA, the UK, and Australia, which were all published in English. Second, the TCE intervention differed greatly in terms of exercise type (Tai Chi, Qigong, Wuqinxi), duration (6–24 weeks), frequency (2–5 times/week), and control group. In the future, a detailed categorization of different types of TCE programs and controls will be required. Third, most of the included RCTs did not adopt blind methods (subject, therapist, and assessor blinding), which might lead to biased subjective expectations and exaggerate the research findings. Finally, most RCTs employed followed up for few months only, so the long-term efficacy of TCE in patients with CLBP remains unclear.