WHAT IS ALREADY KNOWN ON THIS TOPIC

WHAT THIS STUDY ADDS

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

The UR adhered to the Preferred Reporting Items for Systematic Reviews and Meta-Analysis guidelines.38 Furthermore, the UR was officially registered in PROSPERO (CRD42024572146).

To identify systematic reviews and meta-analyses of RCTs on the association between exercise and multiple health outcomes among individuals with cancer, two investigators independently conducted a comprehensive search in PubMed, Embase, Cochrane and Web of Science databases from inception to 23 July 2024. Additionally, the references of eligible articles were manually scrutinised. The detailed search strategy is displayed in online supplemental table 1.

Two reviewers independently screened titles, abstracts, and full texts to ensure compliance with the inclusion criteria. Any differences were addressed by a third reviewer. Studies were selected based on the PICOS (Population, Intervention, Comparison, Outcome, Study design) framework (figure 1), and details can be found in online supplemental table 2.

Eight trained investigators worked in pairs to independently extract information from each eligible study. Any disparities were adjudicated by a senior investigator. Extracted details included the first author, year of publication, study design, number of cases, total population, intervention, comparison, health outcomes, mean (SD) and effect sizes with 95% CIs.

The quality of systematic reviews and meta-analyses was evaluated using the MeaSurement Tool to Assess systematic Reviews (AMSTAR) checklist, with 1 point awarded for each 'yes' answer (out of 11 items).39 The AMSTAR score was classified as high (8–11), moderate (4–7) and low (0–3).40 41 Three investigators independently assessed the quality of the reviews, with discrepancies resolved by a senior investigator.

Using the Grading of Recommendations, Assessment, Development and Evaluations (GRADE), 12 trained investigators independently assessed the certainty of evidence as high, moderate, low or very low.42 Evidence from RCTs begins as high certainty but can be downgraded based on criteria, including risk of bias, indirectness, inconsistency, imprecision and publication bias.42 43

We categorised effect sizes by intervention, comparator, population and outcomes to get a list of unique associations. Using a random-effects model, we recalculated the pooled effect size with 95% CI for associations identified in eligible meta-analyses.44 45 A P value <0.05 was deemed statistically significant. Heterogeneity was evaluated using the I2 statistic. Values exceeding 50% indicated moderate heterogeneity, and values surpassing 75% suggest large heterogeneity.46 We interpreted the standardised mean difference (SMD) analysis using Cohen’s d effect size criteria: a value of 0.2 indicates a small effect, 0.5 indicates a medium effect and 0.8 indicates a large effect.47

We conducted sensitivity analyses for significant associations with moderate-to-high certainty, excluding studies with small sample sizes (25th percentile) or high risk of bias (based on quality assessments). If no quality assessment was available or lacked specific methods and scores, we re-evaluated the primary studies included in the meta-analysis using Cochrane criteria.48–50 Additionally, we assessed the certainty of evidence from the meta-analyses excluded owing to overlapping data, ensuring whether their results aligned with main analysis. All statistical analyses were conducted using STATA version 17.

Our study examined the impact of exercise on health outcomes among individuals with cancer, across all genders, socioeconomic levels and marginalised communities. The author team, comprising 13 women and 7 men from various disciplines (obstetrics, gynaecology, epidemiology, clinical research, cardiology, and clinical laboratories), included scholars at different stages of their careers. Data collection followed a standardised method, ensuring research integrity across diverse cancer populations. Our methodology remained consistent, regardless of regional, educational or socioeconomic differences.

We identified 3070 records from four databases. After removing duplicates and screening, 80 studies with 485 associations were included in our UR (figure 2). The excluded records during the process of full-text screening are listed in online supplemental table 3.

A total of 485 associations published between 2012 and 2024 are shown in online supplemental tables 4 and 5. Among these associations, the types of exercise included mind–body exercise (138, 28.5%), AE and RE (48, 9.9%), high-intensity interval training (HIIT) (18, 3.7%) and other exercises (281, 57.9%). Additionally, these associations were explored in people diagnosed with various types of cancer, including breast (244, 50.3%), digestive system (20, 4.1%), haematological malignancies (13, 2.7%), lung (47, 9.7%), prostate (12, 2.5%) and others (149, 30.7%).

Based on the random-effects model, 260 (53.6%) and 87 (17.9%) associations respectively reached statistical significance at p<0.05 and p<10⁻³. Additionally, 267 (55.1%), 95 (19.6%) and 123 (25.4%) associations showed low, moderate, and high heterogeneity, respectively (online supplemental table 5).

According to the AMSTAR, 66 (82.5%) and 14 (17.5%) of 80 analyses were deemed to be of high and moderate quality, respectively (figure 3).

Approximately 81 (16.7%) and 152 (31.3%) of the meta-analyses remained high and moderate certainty, respectively. Additionally, the remaining 177 (36.5%) and 75 (15.5%) associations were classified as having low or very low certainty, respectively (online supplemental table 6).

We summarised 55 (11.3%) associations between exercise and cancer-related adverse reactions and prognosis among people with cancer, which had adverse effects (eg, CIPN, cardiac toxicity, etc; 31/485, 6.4%) and prognosis (eg, mortality, postoperative complications, etc; 24/485, 4.9%). Nine (1.9%) or 485 associations exhibited high certainty, while 16 (3.3%) demonstrated moderate certainty (online supplemental figure 1a).

We found 67 (13.8%) associations between exercise and body and strength assessment, including body mass index (BMI) (7/485, 1.4%), body composition (25, 5.2%) and biomarkers (35, 7.2%); 12 (2.5%) and 18 (3.7%) associations showed high and moderate certainty, respectively (online supplemental figure 1b).

A total of 172 (35.5%) associations were summarised, including anxiety (21/485, 4.3%), depression (28, 5.8%), sleep problems (47, 9.7%), QoL (63, 13.0%) and other emotional status (13, 2.7%); 32 (6.6%) and 45 (9.3%) associations showed high and moderate certainty, respectively (online supplemental figure 1c).

There were 191 (39.4%) associations evaluating exercise and pain (10/485, 2.1%), physical health (10, 2.1%), social function (15, 3.1%), muscular strength (15, 3.1%), fatigue (66, 13.6%) and physical function (75, 15.5%); 28/485 (5.8%) and 73 (15.1%) associations exhibited high and moderate certainty, respectively (online supplemental figure 1d).

Several sensitivity analyses of significant associations with certainty greater than moderate according to GRADE were conducted. Excluding studies with a high risk of bias effectively mitigated bias.42 Consequently, four associations were upgraded from moderate to high certainty: supervised exercise improved QoL,88 exercise during treatment enhanced physical function,76 HIIT alleviated cancer-related fatigue82 and programmes that integrated yoga or focused on spiritual growth alongside physical activity alleviated symptoms of depression (online supplemental table 7).74 Compared with the largest trial in the meta-analysis, smaller and medium-sized trials exhibited greater treatment effects.89–92 After excluding RCTs with small sample sizes (below the 25th percentile), the evidence level for five associations was downgraded from moderate to low. The involved outcomes included cognitive function16 and fatigue (online supplemental table 8).85 Additionally, 21 associations were excluded owing to overlap. Certainty was upgraded for two associations.52 85 Seven associations were downgraded,76 79 93–97 while 12 associations maintained their original certainty status (online supplemental table 9).55 63 76 79 96 98 99 Finally, these associations excluded owing to insufficient data for quantitative synthesis are summarised in online supplemental table 10, showing general alignment with the main findings.

Cancer surgery often leads to high postoperative complication rates, hindering recovery and increasing healthcare costs.100–102 Preoperative interventions, including exercise, nutrition and psychological support, have gained attention for optimising health and reducing complications.103–107 The findings confirmed that preoperative AE and RE moderately reduced pain and dyspnoea in patients with lung cancer undergoing surgery, shortened hospital stays and decreased complication rates.55 56

Chemotherapy and targeted therapies can cause cognitive impairment, peripheral neuropathy and cardiotoxicity.108–110 Exercise, with its anti-inflammatory and neuroprotective properties,111 112 improves cerebral perfusion and supports the development of the prefrontal cortex and hippocampus.113 114 Our findings confirmed that AE and RE moderately enhance cognitive function in patients with BC undergoing haematopoietic stem cell transplantation, aligning with prior research.115–117 Mind–body exercises such as yoga, Tai Chi and Qigong improved cognitive function in patients with lung cancer. However, the level of evidence for this association was downgraded from moderate to low in sensitivity analyses after excluding small sample sizes. Therefore, further research with larger sample sizes is necessary to fully elucidate the potential benefits of mind–body exercises on cognitive function in this cancer population.

The impact of exercise on CIPN symptoms is debated, but most studies, including ours, showed significant improvements.22 23 118 119 Multifaceted regimens with nerve gliding and sensory-motor-based strategies effectively alleviated CIPN symptoms and postural stability deficits.53 Neurosensory exercises are emphasised as beneficial for cancer populations with CIPN. Regarding cardiotoxicity, AE mitigated chemotherapy-induced effects by preserving LVEF and improving cardiorespiratory function.52 However, evidence on its effectiveness when combined with chemotherapy is mixed, with some studies showing no impact on LVEF but reduced strain damage.26 27 Further research is needed to clarify the role of exercise in preventing cardiotoxicity.

Individuals undergoing treatment for breast or prostate cancer, often experience weight gain or obesity.120–122 Obesity disrupts physiological processes, creating a tumor-promoting microenvironment.123 Our findings indicated that varied exercises (AE, combined exercise, mind–body exercise and RE) reduced obesity, lowered body fat and improved lean body mass. It is important to highlight those studies exclusively comprising people with stage II+ BC, where a slight increase in BMI was observed following supervised exercise interventions. However, when studies included people with various types of cancer who underwent Tai Chi interventions, BMI declined. This discrepancy might be partly attributed to differences in exercise modalities. Supervised programmes may include resistance training, which could increase muscle mass and raise BMI,124 whereas Tai Chi, a low-impact exercise, may promote weight loss.125 Sensitivity analysis focusing on high-quality RCTs confirmed that exercise significantly lowers BMI (MD=−0.60, 95% CI −1.16 to –0.03) and renders the BMI increase non-significant (SMD=0.12, 95% CI −0.09 to 0.33). Thus, initial BMI increases should be interpreted cautiously, emphasising the need for more high-quality studies and tailored approaches to exercise.

Clinical research has shown that the IGF-1/IGF-1R axis and elevated IGF-1 levels are associated with cancer progression,126–128 poor prognoses129 and increased all-cause mortality.129 Insulin and IGF-1 activate the PI3K/Akt/mTOR and Ras/Raf/MAPK signalling pathways and promote tumour proliferation. Our findings indicated that AE, RE and mind–body exercises reduced serum concentrations of IGF-1 and IGFBP-1 in BC survivors. CRP is a well-established marker of inflammation in cancer,130 and raised CRP levels are associated with advanced disease stages in various types of cancer.131–134 Our findings indicated that a combination of AE and RE was particularly effective in reducing serum insulin and CRP levels in BC survivors.62 Moderate-intensity AE significantly reduced CRP levels in people with stage III colon cancer, but showed no substantial effect in stage I–II colon cancer.135 Similarly, individualised exercise programmes effectively lowered CRP levels in people with lung cancer undergoing chemotherapy,136 whereas no significant impact was observed in people with ovarian cancer.137 These findings highlight the need for future research to delineate the variations across different types of cancer and to comprehensively investigate the mechanisms by which exercise regulates these biomarkers, as well as their potential clinical implications.

Cancer diagnosis and treatment often lead to psychological distress, with moderate to severe depression affecting 15% of people, and mild depression and anxiety at rates of 20% and 10%, respectively.138 139 Long-term survivors also face heightened risks of depression and anxiety.140 Our findings suggested that taking part in exercise greatly enhanced sleep quality, reduced insomnia and alleviated anxiety and depression in people with cancer, thus improving their overall QoL. Specifically, yoga and comprehensive exercise have been shown to positively affect sleep quality,66 67 while Tai Chi and Qigong were effective in reducing anxiety.71 Interventions combining yoga or spiritual growth with physical activity, as well as exercise post-radiotherapy or post-diagnosis, are supported by moderate evidence indicating their potential to alleviate depression.72–74 Notably, exercise improved QoL across various cancer types, including BC, lung cancer, and among elderly people with CRC undergoing chemotherapy.68 69 Additionally, yoga was found to enhance short-term health-related QoL in BC.65 Our findings highlighted the significant benefits of mind–body exercises (Tai Chi, yoga and Qigong) in improving sleep, addressing emotional challenges and enhancing QoL in people with cancer. This provides a valuable complement to existing guidelines, which mainly emphasise moderate to high-intensity AE and RE.8 9 141 142

Up to 80% of cancer populations experience significant VO2_peak impairment,143 which is strongly linked to higher treatment-related toxicity,144 poor QoL,145 fatigue146 and increased mortality.143 147 Exercise therapy has been shown to significantly improve VO2_peak, especially in patients with BC.19 Our study showed that exercising during chemotherapy resulted in greater VO2_peak improvements compared with post-chemotherapy exercise, particularly in people with breast, colon and other cancers. These findings highlighte the critical importance of maintaining an exercise regimen throughout the entire chemotherapy period.

HIIT, AE and RE significantly improved body strength and aerobic capacity, with HIIT notably alleviating cancer-related fatigue and pain. These findings aligned with the conclusions of the OptiTrain Trial,148 149 which confirmed that RE and HIIT during chemotherapy improved muscle strength, reduced pain sensitivity and prevented cardiopulmonary decline in people with BC. Mind–body exercises like yoga, aerobic resistance interval training and Tai Chi also enhanced physical functioning,16 including in those with advanced stage disease.57 Tai Chi and progressive resistance training were particularly effective for bolstering lower-body strength.150 Tai Chi improved shoulder function in people with BC within 3 weeks and effectively reduced pain and enhanced arm strength by 12 weeks.81 Social support plays a vital role in stress management and trauma recovery for those with cancer.151 152 Given that cancer often reduces social interactions, our research showed that regular exercise improves social functionality, increases the likelihood of returning to work, and notably, yoga significantly enhances overall social well-being.

We conducted a comprehensive and up-to-date assessment of systematic reviews and meta-analyses on the health outcomes of exercise in people with cancer. A thorough search across authoritative databases was followed by independent screening, data extraction, recalculations of effect sizes and methodological quality evaluations using AMSTAR criteria, with sensitivity analyses ensuring robust findings. Only RCT-based meta-analyses were included, enhancing reliability and accuracy, with 49% of associations rated as high or moderate certainty based on GRADE guidelines. Our study encompasses a broad spectrum of health outcomes, diverse cancer types and exercise interventions, including not only aerobic and resistance training but also mind–body practices. This broad and clinically significant approach provides valuable, generalisable insights for improving QoL in cancer populations and informing fitness and healthcare professionals.

Several limitations of our study need to be acknowledged. First, the GRADE assessment indicated heterogeneity (I² > 50%) and imprecision (continuous variables with total population size <400; binary categorical variables with events <300) in various meta-analyses. One potential contributing factor is the small sample sizes—48% of meta-analyses included fewer than five original studies—resulting in inconsistent findings and imprecise estimates. This affected outcomes such as anxiety, adiponectin levels, glucose levels, interleukin-10/interleukin-6 ratios, tumour necrosis factor-α levels and vascular endothelial growth factor expression. Variability in handling diverse effect estimates, despite their conceptual comparability, and differences in cancer types, exercise parameters (eg, intensity, duration, frequency), outcome assessment criteria and follow-up periods further contributed to heterogeneity and low-certainty evidence for certain outcomes like cognitive function, early-to-atrial filling velocity ratio and cardiorespiratory fitness. However, low-certainty evidence does not negate potential associations, particularly as future studies generate more data.

Second, the broad nature of an UR limited our ability to account for specific confounding factors or mediators related to exercise interventions. Most current evidence on the effectiveness of exercise during or after cancer treatment derived from RCTs involving supervised or home-based exercise, primarily focusing on survivors of common cancers (eg, breast and lung cancer). These participants often met strict eligibility criteria—such as age, comorbidities and physical capability—leading to samples that were healthier and more motivated than the broader cancer populations. It is recognised that demographic differences, including treatment side effects, vary among cancer populations by type of malignancy; therefore, more research is needed to enhance specificity within exercise oncology literature to better address a wider range of cancer types and stages.

Not applicable.

Not applicable.