Exercise and survival benefit in cancer patients: evidence from a comprehensive meta-analysis

Abstract

Cancer remains a major global health challenge, and growing evidence suggests that physical activity is a key modifiable factor that may improve survival outcomes in cancer patients. However, a comprehensive, large-scale synthesis of the effects of post-diagnosis physical activity across multiple cancer types remains lacking. This meta-analysis aims to systematically evaluate the association between physical activity and survival in patients diagnosed with breast, lung, prostate, colorectal, and skin cancers. We conducted a comprehensive search in PubMed, Web of Science, Scopus, and Cochrane Library for studies on physical activity and cancer survival. Eligible studies (January 2000-November 2024) included adults (≥ 18 years) with breast, lung, prostate, colorectal, or skin cancer. Only prospective cohort and case-control studies reporting hazard ratios (HRs) with 95% confidence intervals (CIs) for overall or cancer-specific mortality were included, with a minimum sample size of 100 and at least six months of follow-up. Meta-analysis was performed using metaanalysisonline.com, applying random-effects models and assessing heterogeneity via the I² statistic. Sensitivity analyses and publication bias (Egger's test, funnel plots) were evaluated. The meta-analysis included 151 cohorts with almost 1.5 million cancer patients. Post-diagnosis physical activity was associated with significantly lower cancer-specific mortality across all five cancer types. The greatest benefit was observed in breast cancer, with a pooled hazard ratio (HR) of 0.69 (95% CI: 0.63-0.75), followed by prostate cancer (HR: 0.73, 95% CI: 0.62-0.87). Lung cancer patients who engaged in physical activity had a 24% lower risk of cancer-specific death (HR: 0.76, 95% CI: 0.69-0.84), while colorectal cancer patients experienced a similar benefit (HR: 0.71, 95% CI: 0.63-0.80). In skin cancer, physical activity was associated with a non-significant reduction in mortality (HR: 0.86, 95% CI: 0.71-1.05). These findings provide robust evidence supporting the survival benefits of post-diagnosis physical activity in cancer patients, particularly for breast, prostate, lung, and colorectal cancers. The results underscore the potential for incorporating structured physical activity interventions into oncological care to improve long-term patient outcomes.

Keywords: Breast cancer; Colorectal cancer; Lung cancer; Physical activity; Prostate cancer; Skin cancer; Survival.

Fig. 1

The flowchart illustrates the systematic selection process of studies examining the association between physical activity and survival of cancer patients

Fig. 2

Meta-analysis of the association between physical activity and survival of breast cancer patients. The forest plot presents the pooled hazard ratios (HRs) for the association of physical activity with all-cause survival (top) and breast cancer-specific survival (bottom). Each study is represented by a red square, with the size proportional to its weight in the meta-analysis. Horizontal lines indicate 95% confidence intervals (CIs). The pooled HRs, derived from random-effects models, suggest a significant inverse association between physical activity and mortality of breast cancer patients (pooled HR: 0.66, 95% CI: 0.62–0.70). The analysis demonstrates moderate heterogeneity across studies (I² = 49%, p < 0.01). Abbreviations: CI: confidence interval; HR:hazard ratio; IV: inverse variance; SE: standard error

Fig. 3

Funnel plots illustrating the relationship between hazard ratios (HRs) and standard errors for the association between physical activity and all-cause (A, C, E), and cancer-specific survival (B, D, F) in cancer patients. Breast cancer (A, B), lung cancer (C, D) and colorectal cancer studies (E, F) are illustrated. The shape and symmetry of the funnel plots can offer insights into potential publication bias, with asymmetrical plots indicating the possibility of selective reporting or publication of the studies with certain outcomes

Fig. 4

The forest plot summarizes the association between physical activity and all-cause (top) and lung cancer-specific mortality (bottom). Each study is represented by a red square, with its size reflecting the study’s weight in the meta-analysis. Horizontal lines indicate 95% confidence intervals (CIs). The results indicate a significant reduction in both all-cause and lung cancer-specific mortality among physically active individuals (pooled HR: 0.77, 95% CI: 0.72–0.82). Abbreviations: CI: confidence interval; HR: hazard ratio; IV: inverse variance; SE: standard error

Fig. 5

Meta-analysis of the association between physical activity and mortality of prostate cancer patients. The analysis stratified outcomes into all-cause survival (n = 4 studies) and prostate cancer-specific survival (n = 9 studies). The diamond represents the pooled effect size for each subgroup and overall effect, with values less than 1.0 indicating reduced mortality risk with physical activity (pooled hazard ratio (HR): 0.68, 95% confidence interval (CI): 0.6–0.77). Low between-study heterogeneity was observed in both subgroups (I² = 27% and I² = 16%, respectively) and in the pooled analysis (I² = 24%, p = 0.20). Abbreviations: CI: confidence interval; HR: hazard ratio; IV: inverse variance; SE: standard error

Fig. 6

Forest plot showing the association between physical activity and mortality of colorectal cancer patients. The analysis includes two subgroups: all-cause mortality (upper panel, n = 25 studies) and colorectal cancer-specific mortality (lower panel, n = 15 studies). The pooled analysis demonstrates a significant protective effect of physical activity, with a pooled HR of 0.71 (95% CI: 0.67–0.75). Physical activity showed similar beneficial associations on all-cause survival (HR: 0.71, 95% CI: 0.66—0.75) and colorectal cancer-specific survival (HR: 0.71, 95% CI: 0.63—0.80). Moderate heterogeneity was observed across studies (I² = 45%, p < 0.01). Abbreviations: CI: confidence interval; HR: hazard ratio; IV: inverse variance; SE: standard error