Tai Chi for treating cancer-related fatigue: A meta-analysis of randomized controlled trials

Abstract

BACKGROUND

Cancer-related fatigue (CRF) lacks effective pharmacological treatment, with the available options in Western medicine often having limited efficacy and adverse effects. Tai Chi, a traditional Chinese exercise, shows promise in improving CRF.

AIM

To evaluate the clinical efficacy of Tai Chi in alleviating CRF.

METHODS

In this meta-analysis, we reviewed 9 randomized controlled trials (RCTs) retrieved from databases such as PubMed, EMBASE, the Cochrane Library, China National Knowledge Infrastructure, Wanfang Database, and the Chinese Biomedical Literature Database, and published before March 31, 2025. The experimental groups received conventional treatment plus Tai Chi, and the control groups received conventional treatment only, with varying durations. Using random-effects models, we calculated standardized mean differences (SMD) and mean differences with 95% confidence intervals (CI) to assess the effects of CRF. Heterogeneity was evaluated through I² statistics. To assess the robustness of the pooled results, we performed leave-one-out sensitivity analysis by sequentially excluding each study and reconducting the meta-analysis. Publication bias was evaluated through funnel plot inspection, supplemented by quantitative assessments using the trim-and-fill method and Egger’s test.

RESULTS

This study conducted a systematic review and meta-analysis of 9 RCTs (n = 659 cancer patients) and found that Tai Chi significantly improved CRF, enhanced sleep quality, and increased quality of life, with a favorable safety profile. The research provides evidence-based medical support for promoting Tai Chi as an adjunctive therapy for CRF.

CONCLUSION

Results analysis based on the GRADE assessment indicated that Tai Chi significantly alleviated fatigue symptoms in cancer patients (moderate-certainty evidence, SMD = -1.29, 95%CI: -1.72 to -0.85, P < 0.00001), improved sleep quality (moderate-certainty evidence, SMD = -0.45, 95%CI: -0.78 to -0.12, P = 0.007), and enhanced quality of life (low-certainty evidence, SMD = 0.70, 95%CI: 0.23 to 1.16, P = 0.003), suggesting that Tai Chi can serve as an effective adjuvant intervention for CRF.

Key Words: Tai Chi; Cancer-related fatigue; Meta-analysis; Randomized controlled trial; Traditional Chinese Medicine

Core Tip: This meta-analysis offers moderate-certainty evidence indicating that Tai Chi is a safe and effective mind-body intervention for reducing cancer-related fatigue (CRF) and enhancing sleep quality in cancer patients. Potential benefits were also observed regarding quality of life; however, this finding is supported by low-certainty evidence. Overall, these results support the use of Tai Chi as a complementary therapy for alleviating CRF.

INTRODUCTION

Global data on cancer epidemiology indicate that approximately 19.3 million new cancer cases and nearly 10 million cancer-related deaths occurred worldwide in 2020, and that the overall incidence and mortality rates demonstrated a sustained increase[1]. Apart from posing a significant threat to life, cancer leads to numerous complications and symptoms, with cancer-related fatigue (CRF) being one of the most common. The overall incidence of CRF has been reported up to 43%, although it varies considerably depending on cancer type, treatment stage, and assessment methods[2]. CRF primarily stems from the cancer itself or as an adverse effect of its treatment. It is defined as “the experience of physical, emotional, and/or mental exhaustion related to cancer or its treatment. CRF is not the typical tiredness that follows an active or long day. It is a lack of energy that is distressing, does not improve with normal! Amounts of rest or sleep, and disrupts daily life”[3]. CRF is more severe than ordinary fatigue and is not alleviated by rest. It not only affects the physical functions of the patient but may also cause emotional distress and cognitive decline, thereby significantly compromising the patient’s quality of life[4]. The development and progression of CRF depend on various factors, including the nature of the tumor itself and the associated immune responses, the extent of inflammation, metabolic activity, neuroendocrine changes, the presence/absence of hypothalamic-pituitary-adrenal (HPA) axis dysfunction, and genetic factors influencing treatment. Moreover, even after completion of treatment, persistent pathological changes and disruptions in the body's homeostasis may lead to prolonged inflammatory states, contributing to chronic fatigue[5-10]. Exercise has been explicitly recommended by the National Comprehensive Cancer Network (NCCN) guidelines as one of the important non-pharmacological interventions for managing CRF[3]. Tai Chi is a form of traditional Chinese exercise therapy that embodies a profound cultural heritage and unique health principles. The gentle, rhythmic movements of Tai Chi help regulate emotions, unblock meridians, harmonize qi and blood, and balance yin and yang, thereby improving metabolism, enhancing mood, and boosting immune function[11]. Numerous randomized controlled trials (RCTs) have investigated the effects of Tai Chi on CRF. Moreover, a meta-analysis indicates that compared with other forms of qigong and yoga, Tai Chi is the most effective in improving breast cancer patients' CRF and quality of life[12]. However, for more robust conclusions, it is necessary to pool the data and account for the heterogeneity of results by conducting a meta-analysis with random effects and sensitivity analyses. Such a quantitative synthesis significantly elevates the level of evidence[13-15].

Based on this, we conducted a systematic review of relevant domestic and international literature and conducted a meta-analysis of RCTs on the therapeutic role of Tai Chi in CRF. This investigation aimed to evaluate the clinical efficacy of Tai Chi in preventing and managing CRF and to gain insight into the current evidence regarding this topic.

MATERIALS AND METHODS

Study protocol

We developed a predefined study protocol before conducting this meta-analysis.

Search strategy

The search strategy formulated for this study was in strict accordance with the PICOS (Population, Intervention, Comparison, Outcome, and Study Design) principles. We conducted a comprehensive literature search of multiple databases, namely, PubMed, EMBASE, the Cochrane Library, China National Knowledge Infrastructure, Wanfang Database, and the Chinese Biomedical Literature Database, to identify all relevant studies examining the effects of Tai Chi on CRF. The search period extended from the inception of each database before March 31, 2025. The search strategy employed a combination of Medical Subject Headings and free-text terms using Boolean operators (AND/OR/NOT) to construct search queries, which were adapted according to the specific characteristics of each database to ensure both comprehensiveness and accuracy. The following search terms were used: “Tai Chi” or “Taijiquan” AND “cancer” or “neoplasm” or “malignant” or “tumor” or “radiotherapy” or “chemotherapy” or “immunotherapy” AND “cancer-related fatigue” or “CRF” or “fatigue” or “cancer-induced fatigue” AND “randomized controlled trial” or “RCT” or “randomized controlled” or “randomized”. The computerized search was supplemented by manual searching, and the references and citations of the retrieved articles were screened to ensure comprehensive coverage. It is worth noting that the literature search in this study was limited to research findings published in English and Chinese. English literature searches employed a structured search strategy (Tables 1, 2, and 3), while Chinese literature searches were based on relevant Chinese databases and search terms. To enhance the comprehensiveness of the literature coverage, we conducted a manual review and screening of the references and citations in the retrieved articles.

Table 1 Search strategy for PubMed.

Step	Search strategy
1	Tai Ji[Mesh] OR Tai Chi[Title/Abstract] OR Taijiquan[Title/Abstract] OR Tai Ji Quan[Title/Abstract]
2	Neoplasms[Mesh] OR Cancer[Title/Abstract] OR Tumor[Title/ Abstract] OR malignant[Title/Abstract] OR chemotherapy[Title/Abstract] OR radiotherapy[Title/Abstract] OR immunotherapy [Title/Abstract]
3	fatigue[MeSH] OR cancer-related fatigue[Title/Abstract] OR CRF [Title/Abstract] OR exhaust[Title/Abstract] OR tired[Title/Abstract]
4	randomized controlled trial[Mesh] OR random[Title/Abstract] OR trial[Title/Abstract] OR RCT[Title/Abstract]
5	1 AND 2 AND 3 AND 4

Table 2 Search strategy for EMBASE.

Step	Search strategy
1	'tai chi chuan': Ab,ti OR 'tai ji': Ab,ti OR 'taiji quan': Ab,ti OR taijiquan: Ab,ti OR 'tai chi': Ab,ti
2	'neoplasia': Ab,ti OR neoplasms: Ab,ti OR tumor: Ab,ti OR tumour: Ab,ti OR neoplasm: Ab,ti OR cancer: Ab,ti OR 'malignant neoplasm': Ab,ti OR malignancy: Ab,ti OR malignancies: Ab,ti OR chemotherapy: Ab,ti OR radiotherapy: Ab,ti OR immunotherapy: Ab,ti
3	'fatigue': Ab,ti OR 'cancer-related fatigue': Ab,ti OR crf:ab,ti OR lassitude: Ab,ti OR tiredness: Ab,ti OR exhaustion: Ab,ti OR weariness: Ab,ti
4	'randomized controlled trial': Ab,ti OR 'controlled trial, randomized': Ab,ti OR 'randomised controlled study': Ab,ti OR 'randomised controlled trial': Ab,ti OR 'randomized controlled study': Ab,ti OR 'trial, randomized controlled': Ab,ti OR 'randomized controlled trial': Ab,ti
5	1 AND 2 AND 3 AND 4

Table 3 Search strategy for the Cochrane library.

Step	Search strategy
1	(Tai Chi): Ti,ab,kw OR (Taiji): Ti,ab,kw OR (T'ai Chi): Ti,ab,kw OR (Chi, Tai): Ti,ab,kw OR (Tai Chi Chuan): Ti,ab,kw
2	(Ji Quan, Tai): Ti,ab,kw OR (Taijiquan): Ti,ab,kw OR (Tai Ji Quan): Ti,ab,kw OR (Tai-ji): Ti,ab,kw OR (Quan, Tai Ji): Ti,ab,kw
3	1 OR 2
4	(Neoplasms): Ti,ab,kw OR (Neoplasm): Ti,ab,kw OR (Tumor): Ti,ab,kw OR (Tumors): Ti,ab,kw OR (Neoplasias): Ti,ab,kw
5	(Neoplasia): Ti,ab,kw OR (Malignancy): Ti,ab,kw OR (Malignant): Ti,ab,kw OR (Malignant Neoplasms): Ti,ab,kw OR (Malignancies): Ti,ab,kw
6	(Malignant Neoplasm): Ti,ab,kw OR (cancer): Ti,ab,kw OR (cancers): Ti,ab, kw
7	4 OR 5 OR 6
8	(fatigue): Ti,ab,kw OR (cancer-related fatigue): Ti,ab,kw OR (CRF): Ti,ab,kw OR (lassitude): Ti,ab,kw OR (tiredness): Ti,ab,kw
9	(exhaustion): Ti,ab,kw OR (weariness): Ti,ab,kw
10	8 OR 9
11	(randomized controlled trial): Ti,ab,kw OR (controlled clinical trial): Ti,ab,kw OR (randomized): Ti,ab,kw OR (placebo): Ti,ab,kw OR (random): Ti,ab,kw
12	(trial): Ti,ab,kw OR (group): Ti,ab,kw
13	11 OR 12
14	3 AND 7 AND 10 AND 13

Inclusion criteria

Study design: RCTs with two or more groups, irrespective of the presence/absence of blinding (patient and/or assessor).

Study subjects: Cancer patients diagnosed based on pathologic and imaging findings, with CRF, irrespective of gender, age, region, cancer type, CRF severity, or treatment modality.

Study groups: A control group that received conventional cancer treatment (e.g., radiotherapy, chemotherapy, psychological support, health education, or routine care) alone, and an intervention group in which Tai Chi was administered in addition to conventional treatment.

Outcome measures: (1) Primary outcomes: Fatigue severity is assessed using standardized scales, including the Brief Fatigue Inventory (BFI), Revised Piper Fatigue Scale (RPFS), RPFS-Chinese Version, and Multidimensional Fatigue Symptom Inventory-Short Form (MFSI-SF); (2) Secondary outcomes: Sleep quality, as per Pittsburgh Sleep Quality Index (PSQI) or Self-Rating Scale of Sleep (SRSS); Quality of Life assessed by Functional Assessment of Cancer Therapy-General (FACT-G), FACT-Breast (FACT-B), and WHO Quality of Life-BREF (WHOQOL-BREF); and (3) Safety outcomes: Incidence of adverse events.

Exclusion criteria

Studies were excluded from the meta-analysis if they: (1) Included patients with comorbid conditions that may interfere with study outcomes; (2) Included interventions that combined acupuncture, massage, or other traditional Chinese therapies that may confound the independent evaluation of Tai Chi effects; (3) Were duplicate publications or studies reusing data; (4) Had incomplete or unextractable data; (5) Were in the form of non-original research literature such as conference papers, announcements, commentaries, or reviews; or (6) Were preprints.

Literature screening and data extraction

Two researchers independently screened the titles and abstracts of the articles identified after applying the inclusion and exclusion criteria, which was followed by a full-text assessment. The selection process was recorded using a PRISMA flowchart. Data extraction was performed with a systematic collection of data regarding basic study characteristics, participant information, detailed descriptions of interventions, outcome measure data, and methodological quality-related information. The data was extracted independently by two researchers using a pre-designed standardized data collection form. Discrepancies, if any, were resolved through mutual discussion or by consulting a third researcher to ensure the accuracy and consistency of the extracted data.

Literature quality assessment

All studies included in this analysis were subjected to a rigorous risk-of-bias evaluation using the quality assessment criteria outlined in the Cochrane Handbook for Systematic Reviews (Version 5.1.0)[16]. The assessment covered the following five key domains: Accuracy of the method of random sequence generation; presence/absence of allocation concealment; presence/absence of blinding of researchers and/or participants; completeness of outcome data; and presence/absence of selective reporting of results. To ensure objectivity and accuracy, the quality assessment of each included study was independently performed by two researchers who had received systematic training. A standardized assessment form was used to classify the risk of bias for each domain as "low risk", "high risk", or "unclear risk". In cases of disagreement, consensus was reached through group discussion or consultation with an expert. Finally, RevMan 5.4 software was used to generate a risk-of-bias summary graph and a risk graph, which provided a visual representation of the quality assessment results for each study.

Statistical analysis

In this study, we conducted a meta-analysis using both Stata 17.0 and RevMan 5.4 in a complementary manner. To eliminate scale differences and ensure comparability of results, the standardized mean difference (SMD) with 95% confidence intervals (95%CI) was determined for parameters that were evaluated using different measurement tools in the included studies. We used RevMan 5.4 to assess Cochrane’s risk of bias, which included generating risk-of-bias graphs and summaries, along with forest plots. Using Stata 17.0, we performed sensitivity and publication bias analyses: Publication bias was assessed using funnel plots, Egger’s test, and the trim-and-fill method, while sensitivity analysis was conducted using a leave-one-out approach. If all the included studies used the same measurement tool with consistent units for a given parameter, the MD and its 95%CI were used instead to obtain more clinically interpretable results. Heterogeneity among studies was assessed using Cochrane’s Q test (significance level α = 0.1) with the I² statistic. An I² of < 50% indicated low-to-moderate heterogeneity, and a fixed-effects model was applied for data synthesis. If I² was ≥ 50%, substantial heterogeneity was assumed, and a random-effects model was used, followed by further exploration of the source of heterogeneity[16]. The steps for heterogeneity analysis were as follows: Verify data entry accuracy to rule out human error and conduct subgroup analyses through a literature review to identify potential sources of heterogeneity. To assess the robustness of the pooled results, a leave-one-out sensitivity analysis was performed, where each study was sequentially excluded, and the meta-analysis was re-run. If the direction of the effect or statistical significance remained largely unchanged, the findings were considered reliable.

GRADE evidence quality assessment

The GRADEpro GDT software was used to systematically assess the quality of evidence based on GRADE standards. By evaluating five key downgrading factors—risk of bias, inconsistency, indirectness, imprecision, and publication bias—the quality of research evidence was classified into four levels: High, moderate, low, and very low. The higher the evidence level, the stronger the reliability of the conclusion[17].

RESULTS

Literature search results

The protocol and results of the literature screening process are detailed in Figure 1. The systematic searches of Chinese and English databases initially identified 139 potentially relevant articles. After eliminating duplicate publications, screening the titles and abstracts, and conducting full-text reviews, 9 studies[18-26] were finally determined to meet the predefined criteria for inclusion in further analysis.

Figure 1 Flowchart of literature screening.

Basic characteristics of included studies

We included a total of 9 RCTs[18-26] published between 2013 and 2022. The included studies encompassed a total of 659 patients with CRF, of which 328 patients received only conventional treatment (control group), and 331 received Tai Chi in addition to conventional treatment. Thus, the sample size between the two groups was balanced. The study participants had different types of cancers, namely breast cancer (n = 4)[18,20,22,23], lung cancer (n = 3)[19,21,24], gastric cancer (n = 1)[25], and nasopharyngeal carcinoma (n = 1)[26], thereby indicating that Tai Chi intervention has broad application in fatigue management across different cancer types. Treatment efficacy for CRF was primarily assessed using standardized assessment tools such as the BFI, RPFS, Cancer fatigue scale (CFS), and MFSI-SF scales. The baseline characteristics of the included studies were comparable (Table 4).

Table 4 Basic characteristics of the included literatures.

Ref.	Research object	Sample size		Intervention measure		Intervention cycle	Intervener	Outcome measures
		Treatment	Control	Treatment	Control
Chen[18] (2021)	Breast cancer patients after chemotherapy	26	26	Simplified 24-form Tai Chi	Treatment and routine. Training and exercise	16 weeks, 3 times a week, 60 minutes each time	Professional teachers	CFS body composition, blood lipids
Ge and Shentu[19] (2019)	Lung cancer patients undergoing chemotherapy	39	39	Simplified 24-form Tai Chi	Treatment and routine training	6 weeks, twice a day, 30 minutes each time	Full-time nurses	RPFS, FACT-G, T lymphocyte subset
Han et al[20] (2019)	Breast cancer patients undergoing chemotherapy	20	21	Eight forms of Tai chi	Treatment and routine training	12 weeks, twice a day, 5 days a week	Researchers	RPFS-CV
Jiang et al[21] (2013)	Lung cancer patients	30	30	Simplified 24-form Tai Chi	Treatment and routine training	30 days, twice a day, 30 minutes each time	Full-time nurses	BFI, SRSS
Wang[22] (2017)	Breast cancer patients who underwent modified radical mastectomy and chemotherapy	45	41	Tai Chi	Treatment and routine Training and exercise	Starting from 10 days after the operation, 6 months, twice a day, 20 minutes each time	Researchers	CFS, PSQI, WHOQOL-BREF, SAS, SDS
Yang et al[23] (2022)	Breast cancer patients who underwent modified radical mastectomy and chemotherapy	38	41	SimplifiedTai Chi	Treatment and routine. Training and exercise	12 weeks, twice a day, 5 times a week, 20 minutes each time	Researchers	RPFS-CV, FACT-B, Inflammatory factors
Zhang et al[24] (2016)	Lung cancer patients undergoing chemotherapy	38	36	Eight forms of Tai Chi	Treatment and routine training	12 weeks, 3-4 times a week, 60 minutes each time	Professional teachers	MFSI-SF
Zheng et al[25] (2022)	Postoperative patients with gastric cancer	53	53	Simplified 24-form Tai Chi	Treatment and routine training	12 weeks, 7 days after discharge, 3 times a week, 60 minutes each time	Full-time nurses	RPFS, PSQI
Zhou et al[26] (2018)	Patients with nasopharyngeal carcinoma undergoing chemoradiotherapy	42	41	Simplified 24-form Tai Chi	Treatment and routine training	During chemotherapy, 5 times a week, 60 minutes each time	Researchers	MFSI-SF, Heart rate variability parameters

CFS: Cancer fatigue scale; BFI: Brief fatigue inventory; RPFS: Revised piper fatigue scale; RPFS-CV: Revised piper fatigue scale-Chinese version; MFSI-SF: Multidimensional fatigue symptom inventory-short form; PSQI: Pittsburgh sleep quality index; SRSS: Self-rating scale of sleep; FACT-G: Functional assessment of cancer therapy-general; FACT-B: FACT-breast; WHOQOL-BREF: WHO quality of life-BREF.

Risk-of-bias assessment results for included studies

Among the 9 included studies, six studies[19,20,22-26] provided detailed descriptions of the randomization methods employed. Five of these six studies[19,22-24,26] used simple randomization methods (e.g., random number tables or computer-generated random sequences), while one study[20] used stratified block randomization. In terms of concealment of allocation, three studies[20,24,26] explicitly reported specific implementation methods: Han et al[20] used sealed envelopes, and Zhang et al[24] and Zhou et al[26] delegated allocation to third-party personnel not involved in the recruitment of the study participants. These measures effectively ensured the independence and fairness of the allocation processes adopted in the respective studies. Since Tai Chi is a distinctive form of physical activity, most studies did not clearly describe blinding procedures. However, all studies reported complete outcome data, and there were no instances of selective reporting or other potential sources of bias. The risk-of-bias assessment results are presented in Figures 2 and 3.

Figure 2 Risk-of-bias graph: Authors' judgments about each risk-of-bias item presented as percentages across all included studies.

Figure 3 Risk-of-bias summary: Review authors' judgements about each risk-of-bias item for each included study.

Meta-analysis results

CRF: The 9 RCTs included in the current meta-analysis used different fatigue assessment scales (Table 5)[27-31]. Therefore, we determined the SMD for effect size pooling. Heterogeneity testing revealed substantial heterogeneity (I² = 84%) across the studies, which prompted the use of a random-effects model for analysis. The results of the pooled analysis indicated that Tai Chi intervention was significantly superior to the conventional intervention in improving CRF (SMD = -1.29, 95%CI: -1.72 to -0.85, P < 0.00001) (Figure 4A). These findings suggest that Tai Chi can significantly alleviate fatigue symptoms in cancer patients, with clinically meaningful effects. To further explore potential sources of heterogeneity, we conducted subgroup analyses based on variables such as cancer type, intervention duration, treatment modality, control measures, and assessment scales (Table 6). The results showed that the aforementioned factors exerted varying degrees of influence on the pooled effect sizes, indicating that the high heterogeneity among studies was not randomly generated but rather stemmed from clinical and methodological differences, particularly variations in assessment scales. However, it is worth noting that even within each subgroup, the heterogeneity index (I²) remained consistently high, suggesting that there may be other unidentified sources of variability beyond the aforementioned factors. Given the currently small sample sizes in both the overall and subgroup analyses, the robustness of these conclusions requires validation through larger-scale or multicenter studies. Additionally, four studies[19,20,23,25] assessed fatigue severity in the behavioral, affective, sensory, and cognitive dimensions, and their results indicated that patients in the Tai Chi group had significantly lower fatigue scores than those receiving conventional treatment (MD = -1.19, 95%CI: -1.37 to -1.02, P < 0.00001) (Figure 5). These results together suggest that the positive effects of Tai Chi are experienced across multiple dimensions of fatigue-related symptoms.

Figure 4 Forest map for meta-analysis. A: Tai Chi intervention in cancer-related fatigue; B: Tai Chi intervention in improving sleep quality; C: Tai Chi intervention in improving quality of life.

Figure 5 Subgroup analysis of forest maps based on revised piper fatigue scale/piper fatigue scale.

Table 5 Psychometric properties and applicability of multidimensional fatigue Scales in cancer-related fatigue assessment.

Scale	Items	Score range	Domains assessed	Reliability (α)	Characteristics
CFS[27-31]	15	0-60	Physical, affective, and cognitive	0.88	Describes the fatigue of current situation
BFI[27-31]	9	0-10 per subscale	Fatigue intensity	0.96	Examines fatigue in the previous 24 hours
RPFS[27-31]	22	0-10 per subscale	Behavioural/intensity, cognitive, affective and sensory	0.97	Multidimensional quantitative assessment reflects the complexity of patient fatigue and its impact on life. Addresses CRF's dual physical-cognitive pathology
RPFS-CV[27-31]	22	0-10 per subscale		0.91
MFSI-SF[27-31]	30	0-4 per subscale	General, physical, Emotional, mental and Vigor	0.87-0.96	Ease of use and scoring are of a moderate level

CRF: Cancer-related fatigue; CFS: Cancer fatigue scale; BFI: Brief fatigue inventory; RPFS: Revised piper fatigue scale; RPFS-CV: Revised piper fatigue scale-Chinese version; MFSI-SF: Multidimensional fatigue symptom inventory-short form.

Table 6 A narrative assessment of the efficacy of Tai Chi for cancer-related fatigue: Analysis of sources of heterogeneity.

Analysis dimensions	Number of studies	Ref.	SMD (effect size)	I2	Effect direction and consistency	Notes
Grouped by cancer type
Breast cancer	4	Chen[18] (2021)	-1.70 (-2.51, -0.89)	87%	Negative and consistent	Statistically significant negative effect
		Han et al[20] (2019)
		Wang[22] (2017)
		Yang et al[23] (2022)
Lung cancer	3	Ge and Shentu[19] (2019)	-0.98 (-1.44, -0.53)	79%	Negative and consistent	Statistically significant negative effect
		Jiang et al[21] (2013)
		Zhang et al[24] (2016)
Gastric cancer	1	Zheng et al[25] (2022)				Single-study estimate
Nasopharyngeal carcinoma	1	Zhou et al[26] (2018)				Single-study estimate
Grouped by treatment methods
Undergoing chemotherapy	6	Chen[18] (2021)	-1.51 (-2.15, -0.87)	88%	Negative and consistent	Statistically significant negative effect
		Ge and Shentu[19] (2019)
		Han et al[20] (2019)
		Wang[22] (2017)
		Yang et al[23] (2022)
		Zhang et al[24] (2016)
Undergoing chemoradiotherapy	1	Zhou et al[26] (2018)				Single-study estimate
Routine treatment	2	Jiang et al[21] (2013)	-0.73 (-1.28, -0.19)	64%	Negative and consistent	Statistically significant negative effect
		Zheng et al[25] (2022)
Grouped by intervention time
≤ 4 weeks	2	Jiang et al[21] (2013)	-0.48 (-0.81, -0.15)	0%	Negative and consistent	Statistically significant negative effect
		Wang[22] (2017)
4 < weeks ≤ 8	4	Ge and Shentu[19] (2019)	-0.95 (-1.61, -0.28)	86%	Negative and consistent	Statistically significant negative effect
		Han et al[20] (2019)
		Yang et al[23] (2022)
		Zhou et al[26] (2018)
8 < weeks ≤ 12	5	Han et al[20] (2019)	-1.30 (-1.97, -0.62)	89%	Negative and consistent	Statistically significant negative effect
		Wang[22] (2017)
		Yang et al[23] (2022)
		Zhang et al[24] (2016)
		Zheng et al[25] (2022)
> 12 weeks	2	Chen[18] (2021)	-1.27 (-2.12, -0.42)	79%	Negative and consistent	Statistically significant negative effect
		Wang[22] (2017)
Grouped by comparative measures
The control group receiving conventional rehabilitation training	3	Chen[18] (2021)	-1.76 (-2.85, -0.67)	91%	Negative and consistent	Statistically significant negative effect
		Yang et al[23] (2022)
		Wang[22] (2017)
The control group without conventional rehabilitation training	6	Ge and Shentu[19] (2019)	-1.06 (-1.48, -0.64)	76%	Negative and consistent	Statistically significant negative effect
		Han et al[20] (2019)
		Jiang et al[21] (2013)
		Zhang et al[24] (2016)
		Zheng et al[25] (2022)
		Zhou et al[26] (2018)
Grouped by assessment scales
CFS	2	Chen[18] (2021)	-3.53 (-4.45, -2.62)	0%	Negative and consistent	Statistically significant negative effect
		Wang[22] (2017)
RPFS/RPFS-CV	4	Ge and Shentu[19] (2019)	-1.26 (-1.52, -1.00)	56%	Negative and consistent	Statistically significant negative effect
		Han et al[20] (2019)
		Yang et al[23] (2022)
		Zheng et al[25] (2022)
MFSI-SF	2	Zhang et al[24] (2016)	-9.42 (-15.63, -3.22)	69%	Negative and consistent	Statistically significant negative effect
		Zhou et al[26] (2018)

CRF: Cancer-related fatigue; CFS: Cancer fatigue scale; BFI: Brief fatigue inventory; RPFS: Revised piper fatigue scale; RPFS-CV: Revised piper fatigue scale-Chinese version; MFSI-SF: Multidimensional fatigue symptom inventory-short form.

Sleep quality

The three included studies[21,22,25] investigated the effects of Tai Chi on sleep quality in cancer patients. One study[25] was excluded from the final analysis as it only reported the number of cases with improved sleep without providing specific sleep score data. Among the two studies included in the final analysis, Wang[22] employed the PSQI for assessment, while Jiang et al[21] utilized the SRSS. Due to the use of different assessment tools, the SMD was adopted to pool effect sizes. Heterogeneity testing revealed low heterogeneity (I² = 0%), prompting the application of a fixed-effects model for meta-analysis. The results demonstrated that the Tai Chi group showed significantly greater improvement in sleep quality compared to the conventional treatment group (SMD = -0.45, 95%CI: -0.78 to -0.12, P = 0.007) (Figure 4B). These findings support the potential benefits of Tai Chi for sleep quality enhancement.

Quality of life

Three studies[19,22,23] respectively used different scales to assess the impact of Tai Chi on patients' quality of life: Ge and Shentu[19] used FACT-G, Wang[22] used WHOQOL-BREF, and Yang et al[23] used FACT-B. Due to the use of different assessment tools, the SMD was employed as the effect size for the meta-analysis. There was high heterogeneity among the studies (I² = 68%), so a random-effects model was used for the meta-analysis. The results showed that the Tai Chi group had significantly better quality of life scores than the control group (SMD = 0.70, 95%CI: 0.23 to 1.16, P = 0.003) (Figure 4C). These results support the potential benefits of Tai Chi on quality of life.

Other objective indicators

Multiple studies have reported the improvement effect of Tai Chi on several objective indicators of CRF patients. Chen's[18] research indicated that Tai Chi exercise could significantly reduce the body fat, BMI, total cholesterol, triglycerides, and low-density lipoprotein levels of patients (all indicators P < 0.05), while increasing muscle mass and metabolic rate; the control group did not show significant improvement, and its triglyceride level even significantly increased (P < 0.05). Ge and Shentu[19] research found that after Tai Chi intervention, the proportion of CD3⁺ T cells and the ratio of CD3⁺/CD4⁺ in the patients of the experimental group were significantly higher than those of the control group (all P < 0.01), while the proportion of CD3⁺CD8⁺ T cells was significantly lower (P < 0.01). Yang et al[23] compared the changes in inflammatory factors between the two groups and found that the levels of interleukin (IL)-1, IL-6, and tumor necrosis factor-α (TNF-α) in the patients of the Tai Chi intervention group decreased significantly over time (P < 0.05); in the control group, IL-6 slightly increased, but the difference was not statistically significant (P > 0.05). In addition, Zhou et al[26] also reported that after chemotherapy or radiotherapy, the standardized low-frequency/high-frequency power ratio of the patients in the Tai Chi group was significantly lower than that of the control group, and this ratio was significantly correlated with the total score of the MFSI-SF. However, due to the limited available data in the current studies, no further comprehensive analysis of the impact of Tai Chi on these indicators was conducted.

Safety evaluation

Regarding safety outcomes, only the studies by Han et al[20] and Yang et al[23] explicitly monitored and reported the absence of adverse events related to the intervention. The remaining trials did not mention whether adverse events were assessed or recorded. As a result, the available data are insufficient to comprehensively evaluate the safety profile of this intervention.

Sensitivity analysis

Sensitivity analysis using the leave-one-out method showed that the direction of the pooled effect size remained consistent (all negative) after excluding any single study, and the range of variation was limited (-1.48 to -0.88), suggesting that the overall results were robust and that no single study had a decisive influence on the conclusions (Figure 6).

Figure 6 Sensitivity analysis of the methodological quality of the included studies.

Publication bias

Given the limited number of studies included and the relatively small sample size in this research, multiple analytical methods were employed to systematically evaluate potential bias in the study results. These methods included generating funnel plots, performing Egger’s test, and applying the trim-and-fill method to estimate and adjust for potentially missing studies. The results indicated no significant publication bias among the included studies (Figures 7, 8, and 9).

Figure 7 Funnel plot for publication bias in the included studies.

Figure 8 Results of the Egger test for publication bias in the included studies.

Figure 9 Trim-and-fill analysis for publication bias in the included studies. A: Filled funnel plot with pseudo 95%CI limits; B: Iteration summary of the linear trimming estimator.

GRADE evidence quality rating results

The certainty of evidence for each outcome was assessed using the GRADEpro GDT online tool. In the absence of significant methodological limitations, RCTs are considered to have high-certainty evidence in the GRADE system. The results showed that, compared with the control group, the experimental group demonstrated a significant advantage in improving CRF (moderate-certainty evidence). In addition, moderate-certainty evidence indicates that the experimental group significantly improved sleep quality, while low-certainty evidence suggests a potential improvement in quality of life (Table 7).

Table 7 GRADE evidence quality rating results.

Quality assessment							No. of patients		Effect	Quality
Number of studies	Study design	Risk of bias	Inconsistency	Indirectness	Imprecision	Other considerations	Tai Chi intervention	Control	Std. mean difference (95%CI)
Cancer-related fatigue
9	Randomized trials	Not serious	Serious1	Not serious	Not serious	None	331	328	SMD -1.29; (-1.72, -0.85)	Moderate
Sleep quality
2	Randomized trials	Not serious	Not serious	Not serious	Serious2	None	75	71	SMD -0.45; (-0.78, -0.12)	Moderate
Quality of life
3	Randomized trials	Not serious	Serious3	Not serious	Serious2	None	122	121	SMD 0.70; (0.23, 1.16)	Low

¹Due to significant heterogeneity among studies (I² = 84%, P < 0.00001), the evidence grade was downgraded by one level owing to serious inconsistency.

²Given the small sample size and limited data, we downgraded the evidence for imprecision by one level.

³Due to significant heterogeneity among studies (I² = 68%, P = 0.003), the evidence grade was downgraded by one level owing to serious inconsistency.

SMD: Standardised mean difference.

DISCUSSION

The findings of this meta-analysis, which included 9 RCTs (n = 659), indicate that Tai Chi significantly improves CRF (moderate evidence), sleep quality (moderate evidence), and quality of life (low-certainty evidence). These results provide evidence-based support for the clinical use of Tai Chi as a non-pharmacological intervention in cancer management. Notably, the observed improvement in fatigue may be closely linked to the enhancement of sleep quality. This connection is further supported by genetic evidence from Liu et al[32], who explored the causal relationship between sleep duration and depression. Their study showed that improving sleep can reduce depression to a certain extent and alleviate fatigue. This finding supports the role of Tai Chi in relieving fatigue symptoms in cancer patients, suggesting that it may achieve secondary benefits by improving sleep quality.

The development of CRF involves complex multisystem interactions. At the molecular level, its pathophysiologic mechanisms primarily revolve around three key pathways: Inflammatory response, neuroendocrine dysfunction, and mitochondrial impairment[33]. First, the persistent stimulation of the tumor microenvironment and anticancer treatments (including chemotherapy, radiotherapy, and targeted therapy) leads to an abnormally activated state of the immune system characterized by a significant increase in circulatory pro-inflammatory cytokines (e.g., IL-6, TNF-α, IL-1β). These inflammatory mediators disrupt tryptophan metabolism by activating the indoleamine 2,3-dioxygenase pathway, causing an imbalance in the serotonin/kynurenine metabolic axis[34]. Additionally, these cytokines directly affect the functioning of the HPA axis, resulting in glucocorticoid receptor dysfunction and abnormal cortisol secretion rhythms, thereby inducing widespread neuroendocrine dysregulation[35]. Apart from this, chemotherapeutic agents (e.g., anthracyclines, platinum-based drugs) can also induce mitochondrial DNA damage, inhibit electron transport chain activity, and interfere with oxidative phosphorylation, thereby significantly compromising the efficiency of cellular energy metabolism. This mitochondrial dysfunction is a key molecular factor involved in the persistence of CRF[36].

Tai Chi exerts multi-target regulatory effects on various pathological processes. With respect to inflammatory response, the therapeutic value of Tai Chi lies in its ability to bidirectionally regulate immune homeostasis. While suppressing the overexpression of pro-inflammatory cytokines (IL-6, TNF-α) by downregulating transcription factors such as NF-κB, it also activates the cholinergic anti-inflammatory pathway via vagus nerve stimulation to promote the release of anti-inflammatory cytokines (e.g., IL-10, TGF-β)[37]. The movements of Tai Chi, which are characterized by "softness, slowness, relaxation, and smoothness", create low-intensity aerobic stimulation through slow, continuous circular motions. The gentle rhythmic and deep-breathing exercises help regulate autonomic nervous system function to enhance parasympathetic activity and thereby alleviate the stress response generated by the HPA axis. Long-term practice of Tai Chi has been shown to reduce baseline cortisol levels, restore its diurnal rhythm, and increase glucocorticoid receptor sensitivity, which in turn helps reduce fatigue, mood disorders, and sleep disturbances caused by neuroendocrine dysfunction[38,39]. Furthermore, Tai Chi may also optimize the negative feedback regulation of the HPA axis by modulating brain-derived neurotrophic factor expression and promoting hippocampal neuroplasticity[40]. Tai Chi alleviates subjective CRF symptoms through psychophysiological synergy. Its meditative components (e.g., mindfulness and breath control) reduce anxiety and depression levels and modulate functional connectivity in the brain's default mode network, thereby diminishing fatigue perception[41,42]. Additionally, the group-based format of Tai Chi practice enhances social support, further improving patients' overall functional status. Tai Chi may alleviate CRF through anti-inflammatory, neuroendocrine-modulating, and psychophysiological mechanisms, though further research is needed to validate its precise molecular and systemic effects.

This study has some limitations. First is the lack of standardized intervention protocols involving Tai Chi, which affects the comparability and reproducibility of the findings across different clinical scenarios. Second is the use of subjective symptom scales to evaluate the efficacy of interventions incorporating Tai Chi; instead, objective biomarkers such as inflammatory cytokine levels, neuroendocrine indicators, heart rate variability, and metabolomic profiles should be integrated into the evaluation systems for a more comprehensive and objective assessment of outcomes. Additionally, the study only focused on the Chinese population, which limits the generalizability of the findings. To further extend the benefits of Tai Chi, efforts should be made to develop personalized intervention strategies for patients. By applying machine learning to analyze clinical characteristics and biomarkers, predictive models for treatment response could be established, enabling precision interventions and outcome assessment.

CONCLUSION

In summary, the meta-analysis reveals that Tai Chi has beneficial effects on CRF, which can be attributed to its complex, multisystem regulatory effects. Future research should focus on two key aspects: Systems biology approaches (e.g., network pharmacology, multi-omics integration) should be employed in basic research to further elucidate its mechanisms of action, and multicenter large-scale studies and real-world research should be conducted to optimize intervention and assessment protocols in clinical practice. This rehabilitation method, embodying Eastern wisdom, should be better integrated into global cancer care to benefit patients worldwide.