Introduction

Chemotherapy-induced cardiotoxicity (CIC) refers to the direct and indirect adverse effects of different chemotherapeutic agents on the cardiovascular system [1]. In particular, the incidence of left ventricular dysfunction among patients treated with certain anticancer drugs, such as doxorubicin at high doses (700 mg/m2), can reach 48%. In contrast, the incidence of myocardial ischemia due to 5-fluorouracil (5-FU) is reported to be as high as 10% [2, 3]. Moreover, 26–93% of patients on arsenic trioxide show prolonged QT interval, and many develop life-threatening ventricular tachyarrhythmias [4]. Besides being a not infrequently occurring event, CIC corresponds to a wide range of adverse events. According to the European Society of Cardiology’s Task Force for Cancer Treatments and Cardiovascular Toxicity, chemotherapy-related cardiovascular complications are classified as myocardial dysfunction and heart failure, coronary artery disease (CAD), arrhythmias, arterial hypertension, thromboembolic disease, peripheral vascular disease, pulmonary hypertension, and pericardial complications [2].

Consequently, different pharmacological and non-pharmacological therapies were investigated as potential preventive approaches against CIC, among them physical exercise, whose efficacy and tolerability were tested by numerous clinical trials with promising results [5, 6]. Several parameters can be used to assess the effects of exercise on cardiac function, such as left ventricular ejection fraction (LVEF), left ventricular end-diastolic volume (LVEDV), left ventricular end-systolic volume (LVESV), and global longitudinal strain (GLS) which are all echocardiographically determined [7]. Besides this, cardiovascular fitness, i.e., peak oxygen uptake (VO2 peak) is also an interesting outcome to evaluate in this context. VO2 peak our primary outcome, is the peak value of oxygen uptake attained during exercise [8]. In a recent meta-analysis, high-intensity interval training positively affected cancer patients' functional performance [6]. Similarly, it was reported that exercise training can ameliorate cardiorespiratory fitness following chemotherapy with anthracyclines [9]. Additionally, the randomized controlled trial (RCT) known as The BREXIT Study has demonstrated that exercise can effectively prevent anthracycline-induced functional disability and cardiac impairment [10]. In contrast, another RCT has concluded the lack of feasibility of intensive aerobic training in a significant proportion of patients with metastatic breast cancer receiving chemotherapy [11].

Thus, it is not clear if the current data is sufficient to encourage the use of exercise for patients at risk of CIC, especially since exercise is not currently a part of the recommended standards of care for cancer management [12]. Furthermore, most established cardio-protective exercise abilities were observed in non-cancer populations [5]; therefore, the same effects may not necessarily be seen in cancer survivors.

This creates a solid rationale to extensively examine the findings of the current literature to provide a vigorous assessment of exercise advantages in lowering the risks of cardiovascular events following chemotherapy. Consequently, in the present systematic review and meta-analysis, we explored the quality of evidence that determines exercise's cardiac efficacy and safety in patients receiving chemotherapy. Our work may lead to insightful findings that can have key therapeutic implications.

Methodology

Protocol registration

The PRISMA statement and the Cochrane Handbook for systematic reviews and meta-analyses were followed to conduct this systematic review and meta-analysis [13, 14]. This meta-analysis process has been registered and published in PROSPERO under the following ID: CRD42023460902.

Data sources & search strategy

PubMed (MEDLINE), Scopus, Cochrane Central Register of Controlled Trials (CENTRAL), Web of Science Core Collection, EMBASE, Clinical Trials.gov, and MedRxiv were systematically searched until July 17th, 2023. We modified search terms and keywords for each database, as presented in (Table S1).

Eligibility criteria

We included randomized controlled trials (RCTs) published in English language that followed the following PICO criteria: population (P): patients diagnosed with any type of cancer receiving any cardiotoxic chemotherapeutic agent; intervention (I): any form of supervised aerobic or resistance exercise training irrespective of the exercise duration, frequency and intensity; control (C): usual care without any form of exercise training; and outcomes (O): primary outcome of this review is the VO2 peak. While our secondary outcomes include left ventricular ejection fraction (LVEF) change, change in global longitudinal strain (GLS), cardiac output (CO) (L/min) change, stroke volume (SV) (ml) change, left ventricular end-diastolic volume (LVEDV) (ml) change, left ventricular end-systolic volume (LVESV) (ml) change, E/A ratio change, respiratory exchange ratio (RER) change, resting heart rate (RHR) change, peak heart rate (PHR) change, resting systolic blood pressure (RSBP) (mmHg) change, resting diastolic blood pressure (RDBP) (mmHg) change, and safety outcomes, including the incidence of any adverse events, any serious adverse events, any adverse events leading to withdrawal, and mortality.

Study selection

To perform the review, we used the Covidence web tool. After deleting duplicates, four authors (M.T., M.I., A.N., and H.S.) independently evaluated the obtained records. Four authors (M.T., M.I., A.N., and H.S.) checked the full texts of the records that satisfied the initial eligibility criterion during the full-text screening. Any differences were settled by discussion and agreement with B.A.

Data extraction

We conducted a pilot extraction after retrieving the complete texts of relevant papers in order to prepare the data extraction sheet appropriately. The data extraction sheet, which is structured in Excel (Microsoft, USA), is divided into three sections. The first part included the summary characteristics of the included studies (name of first author, year of publication, country, exercise intensity, intervention frequency (Sessions per week), chemotherapeutic drug, exercise adherence, cancer type, cancer stage, and study design). The second part included the baseline information of the participants (sample size, age, menopausal status, body mass index (BMI), cancer stage, and comorbidities). Finally, the third part included outcomes data as previously described. Four reviewers (M.T., M.I., A.N., and H.S.) were responsible for data extraction. Any differences were settled by discussion and agreement with B.A.

Risk of bias and certainty of evidence

Using the Cochrane RoB2 tool, four reviewers (M.T., M.I., A.N., and H.S.) independently evaluated the quality of the studies [15]. They assessed five domains, including the risk of bias associated with the randomization process, deviation from the intended intervention, missing outcome data, measuring the outcome, and choosing the reported results. Any differences were settled by discussion and agreement with B.A. Two reviewers (M.A. and B.A.) followed the Grading of Recommendations Assessment, Development, and Evaluation (GRADE) criteria [16, 17] to evaluate the certainty of evidence. Any disagreements were resolved through consensus.

Statistical analysis

The RevMan v5.3 software was used for the statistical analysis [18]. We employed the risk ratio (RR) to combine the results of dichotomous outcomes and the mean difference (MD) for continuous outcomes, both with a 95% confidence interval (CI), using the fixed-effects model. However, the random-effects model was used in case of significant heterogeneity. To assess heterogeneity, we utilized the Chi-square and I-square tests, where the Chi-square test establishes if heterogeneity exists, and the I-square test assesses the level of heterogeneity. According to the Cochrane Handbook (chapter nine) [19], we considered an alpha level of less than 0.1 for the Chi-square test to indicate significant heterogeneity, while an I-square more than 75% indicated considerable heterogeneity. When there was significant heterogeneity, sensitivity analysis was used in which we excluded one study in each scenario to detect possible heterogeneity causes.

Trial Sequential Analysis (TSA) was employed to assess the conclusiveness and reliability of the data of the pooled trials and to assess if the sample size of the current meta-analysis was adequate to make solid conclusions regarding the impact of the interventions. When the Z-line on the curve cut both the conventional and trial sequential monitoring boundary (TSMB), we assumed that the intervention's confidence level was conclusive and sufficient and that no additional studies were required. However, if the Z-line does not cut any boundaries, the evidence is insufficient, and further studies are needed [20, 21]. In this meta-analysis, we utilized an alpha error of 0.05, a beta error of 80% power, and a predicted RR reduction of 20% in dichotomous outcomes. Moreover, we made a subgroup analysis based on exercise type (aerobic exercise, restrictive exercise, and combined aerobic and restrictive exercise) and regarding whether the patients had breast cancer only or breast cancer plus other cancers throughout our primary and echocardiographic outcomes to detect possible differences between the subgroups.

Results

Search results and study selection

This literature search from PubMed (MEDLINE), Scopus, Cochrane Central Register of Controlled Trials (CENTRAL), Web of Science Core Collection, EMBASE, Clinical Trials.gov, and MedRxiv yielded a total of 4,446 articles. After duplication removal (n = 1371) and reviewing the title and abstract (n = 3075) for relevance, eighty-six articles were left for full-text screening. Thirteen of these studies met the inclusion criteria for our systematic review and meta-analysis. The PRISMA flow diagram displays the search results and studies selection process (Fig. 1).

Fig. 1
figure 1

PRISMA flow chart of the screening process

Full size image

Characteristics of included studies

This study involves thirteen RCTs [9, 10, 22,23,24,25,26,27,28,29,30,31,32] with a total of 952 patients, diagnosed with various types of cancer undergoing treatment with cardiotoxic chemotherapeutic agents. Among them, 569 (59.77%) patients participated in supervised aerobic or resistance exercise training sessions, whereas 383 (40.23%) did not receive any type of exercise. All the RCTs included the participants with breast cancer except Tsai et al. 2019 [24], which included the patients with Sarcoma hip/thigh, Lymphoma, Multiple myeloma, Osteosarcoma, Hodgkin's disease, and Leukemias as well. Also, in all the included RCTs, participants were delivered moderate to vigorous intensity exercise; however, there were variations in the exercise character, duration, and the number of exercise sessions among the studies. The detailed summary characteristics of the included RCTs and participants’ baseline characteristics are shown in (Table 1 and 2) respectively.

Table 1 Characteristics of the included studies
Full size table
Table 2 Baseline characteristics of the participants included
Full size table

Risk of bias and certainty of evidence

The risk of bias assessment for each outcome is presented in (Fig. 2). Overall, most of the included studies demonstrated a low risk of bias across all assessed domains. Specifically, four studies raised some concerns regarding the risk of bias, primarily stemming from issues related to outcome measurement. Notably, only one study was deemed to have a high risk of bias, primarily due to shortcomings in the randomization process. More details about the authors’ decision are in (Table S2). Certainty of evidence is demonstrated in a GRADE evidence profile (Table 3).

Fig. 2
figure 2

Quality assessment of the risk of bias in the included trials. The upper panel presents a schematic representation of risks (low = green, unclear = yellow, and high = red) for specific types of biases of each of the studies in the review. The lower panel presents risks (low = green, unclear = yellow, and high = red) for the subtypes of biases of the combination of studies included in this review

Full size image
Table 3 GRADE evidence profile
Full size table

Primary outcome

There was a significant difference between exercise and usual care regarding VO2 peak change with (MD: 1.95 with 95% CI [0.59 -3.32], P = 0.005) (Fig. 3-A). The pooled studies were heterogeneous (I2 = 90%, P < 0.00001). Heterogeneity was not resolved by leave-one-out sensitivity analysis (Table S3). TSA showed that the available evidence crossed both the conventional boundary and TSMB, indicating robust conclusions (Fig. 3-B). The subgroup analysis showed a significant difference in exercise type subgroups (P = 0.006) with a significant increase in VO2 peak in the aerobic exercise group (MD: 1.89 with 95% CI [0.23 – 3.55], P = 0.03), and combined exercise group (MD: 2.47 with 95% CI [0.63 – 4.30], P = 0.008). However, there was no difference in the resistant exercise group (MD: 0.10 with 95% CI [-0.16 – 0.37], P = 0.44) (Figure S1). However, test for subgroup analysis was not significant regarding whether the patients had breast cancer only or breast cancer plus other cancers (P = 0.82) (Figure S2).

Fig. 3
figure 3

Forest plot and trial sequential analysis of the primary efficacy outcome (VO2 peak), MD: mean difference, CI: confidence interval

Full size image

Secondary outcomes

Efficacy outcomes

There was no significant difference between exercise and usual care regarding LVEF change (MD: 1.18 with 95% CI [-0.45, 2.81], P = 0.16), GLS change (MD: 0.42 with 95% CI [-0.52, 1.37], P = 0.38), CO change (MD: 0.51 with 95% CI [-1.00, 2.01], P = 0.51), SV change (MD: 2.24 with 95% CI [-9.04, 13.51], P = 0.70), LVEDV change (MD: -2.47 with 95% CI [-8.13, 3.18], P = 0.39), LVESV change (MD: -1.93 with 95% CI [-4.64, 0.78], P = 0.16), E/A ratio change (MD: 0.02 with 95% CI [-0.05, 0.10], P = 0.56) (Fig. 4).

Fig. 4
figure 4

Forest plots of the secondary efficacy outcomes, (1: Left ventricular ejection fraction (LVEF) change, 2: Global longitudinal strain (GLS) change, 3: Stroke volume (SV) change, 4: Left ventricular end-diastolic volume (LVEDV) change, 5: Left ventricular end-systolic volume (LVESV) change, 6: E/A ratio change, and 7: Cardiac output (CO) change), MD: mean difference, CI: confidence interval

Full size image

Moreover, there was no significant difference between exercise and usual care regarding RER change (MD: 0.02 with 95% CI [-0.02, 0.05], P = 0.31) (Figure S3), RHR change (MD: -1.63 with 95% CI [-4.64, 1.39], P = 0.29) (Figure S4), PHR change (MD: 3.45 with 95% CI [-0.35, 7.25], P = 0.08) (Figure S5), RSBP change (MD: -3.32 with 95% CI [-8.79, 2.15], P = 0.23) (Figure S6), RDBP change (MD: -2.47 with 95% CI [-6.39, 1.44], P = 0.22) (Figure S7).

The pooled studies were homogenous in LVEF change (I2 = 39%, P = 0.12), LVEDV change (I2 = 0%, P = 0.72), LVESV change (I2 = 0%, P = 0.90), E/a ratio change (I2 = 0%, P = 0.54), RER change (I2 = 0%, P = 0.75), RHR change (I2 = 0%, P = 0.78), PHR change (I2 = 0%, P = 0.97), RSBP change (I2 = 0%, P = 0.64), and RDBP change (I2 = 0%, P = 0.66). However, pooled studies were heterogeneous in GLS change (I2 = 53%, P = 0.06), CO change (I2 = 97%, P < 0.00001), and SV change (I2 = 94%, P < 0.00001). Regarding GLS change, heterogeneity was best resolved by excluding Antunes et al. 2023 and Jacquinot et al. 2022 (I2 = 19%, P = 0.29), (I2 = 0%, P = 0.44) respectively. Regarding SV change, heterogeneity was best resolved by excluding Foulkes et al. 2023 (The BREXIT) (I2 = 0%, P = 0.43). Regarding CO change, heterogeneity was best resolved by excluding Foulkes et al. 2023 (The BREXIT) (I2 = 45%, P = 0.18) (Table S3). The test of subgroup analysis regarding exercise type was insignificant in all the outcomes. The subgroup analysis can be found in (Figures S8-19). Moreover, test for subgroup analysis was not significant regarding whether the patients had breast cancer only or breast cancer plus other cancers (Figure S20-S23).

Safety outcomes

There was no significant difference between exercise and usual care regarding the incidence of any adverse event (RR: 4.44 with 95% CI [0.47, 41.56], P = 0.19), any serious adverse event (RR: 3.00 with 95% CI [0.14, 65.90], P = 0.49), any adverse event leading to withdrawal (RR: 2.87 with 95% CI [0.79, 10.43], P = 0.11), and all-cause mortality (RR: 0.25 with 95% CI [0.03, 2.22], P = 0.21) (Fig. 5). Pooled studies were heterogenous in any adverse event (I2 = 74%, P = 0.02). However, the pooled studies were homogenous in any adverse event leading to withdrawal (I2 = 0%, P = 0.67) and All-cause mortality (I2 = 0%, P = 0.80). Regarding any adverse event, heterogeneity was best resolved by excluding Foulkes et al. 2023 (The BREXIT) and Kerrigan et al. 2023 (I2 = 45%, P = 0.18), (I2 = 33%, P = 0.22) respectively (Table S3).

Fig. 5
figure 5

Forest plot of the adverse events, RR: risk ratio, CI: confidence interval

Full size image

Discussion

This meta-analysis showed that exercise is an effective enhancer of VO2 peak in chemotherapy patients. Furthermore, compared to usual care, exercise does not elicit any significant improvement in heart function-related parameters, including LVEF, GLS, CO, SV, LVEDV, LVESV, E/A ratio, RER, RHR, PHR, RSBP, and RDBP. Also, exercise-based care was a tolerable approach during chemotherapy that does not expose any additional risks for adverse events, confirming previous results from the oncology population [33,34,35].

VO2 peak refers to the limited value of oxygen uptake/consumption actually achieved during an exercise test (e.g., running on a treadmill). In other words, VO2 peak is the greatest value of the consumed oxygen by an exercising subject independently to his work rate level [36]. Notably, VO2 peak is 30% lower in cancer patients compared to age- and sex-matched healthy individuals who do not practice exercise [37]. Thus, it was shown by Jones et al. to be a strong independent predictor of survival among patients with non-small cell lung cancer. Thus, in these patients, the adjusted hazard ratio of all-cause mortality was 0.64 for a VO2 peak of 0.96–1.29 L.min − 1 and even lower, reaching 0.56 for a VO2 peak of > 1.29 L.min − 1 compared to VO2 peak < 0.96 L.min − 1 [38]. This suggests that a moderate increase in VO2 peak is beneficial to improve prognosis in the oncology population.

Our findings indicate that exercise can protect against chemotherapy-induced drop in VO2 peak, especially since cancer survivors who received neoadjuvant chemotherapy, compared to those who did not receive it, were reported to display a decreased peak VO2 per kg by 23% [39]. It is unclear how exercise would induce this effect; however, several mechanisms seem to be involved. The ability of exercise to reduce body mass index (BMI) during chemotherapy was confirmed by a recent systematic review [40]. Therefore, exercise may improve VO2 peak among chemotherapy patients by decreasing their BMI, as the latter is negatively associated with VO2 peak [41]. Exercise was also found to increase lean mass among cancer survivors, while the absence of exercise favors skeletal muscle loss within the same category [42, 43]. This can contribute to the exercise-induced improvement in cancer-related fatigue in oncology patients as lean mass increase is likely to be accompanied by a VO2 peak increase [44].In line with this, results from animal experiment have demonstrated that in rats receiving doxorubicin (a chemotherapy drug known by its toxic effects on skeletal muscle), preconditioning with exercise had enabled the prevention/minimization of skeletal muscle atrophy, contractile dysfunction, and muscular fatigue [45, 46]. Not just that but endurance exercise was shown to reverse doxorubicin-induced myotoxicity in rats [47]. All this may suggest that VO2 peak can be boosted in exercising oncology patients by a peripheral mechanism through positive effects on muscular growth, strength, metabolic function and recovery which would ultimately ameliorate oxygen uptake at the local level (muscle VO2). Especially that we found no significant benefit of exercise on central (i.e., cardiac) hemodynamics, which makes the peripheral action on skeletal muscle the more likely way to boost VO2 peak after chemotherapy. Moreover, higher systemic inflammation is correlated with lower VO2 peaks among cancer patients [48], and it is well-established that chemotherapy has pro-inflammatory effects. Therefore, exercise may also elevate VO2 peak via its potential to protect cancer survivors from systemic inflammation, particularly chemotherapy [49, 50].

Exercise failed to ameliorate the cardiovascular function of chemotherapy patients, which signifies that training therapy is potentially devoid of substantial protective effects against CIC. The absence of improvement in CO, LVEF, SV, LVEDV, LVESV, GLS, and E/A ratio indicates the inefficacy of exercise in reducing chemotherapy-induced left ventricular dysfunction and heart failure. Moreover, the fact that exercise did not show beneficial chronotropic effects (no changes in RHR and PHR) does not support the protective value of training programs against tachyarrhythmias associated with chemotherapeutic agents. Furthermore, a number of cytotoxic drugs, such as platinum components and alkylating agents, can induce secondary hypertension [51]. The insensibility of RSBP and RDBP to exercise-based therapy shows that the latter may have no notable effects on reducing the susceptibility to chemotherapy-induced hypertension.

It is necessary to determine the safety profile of any intervention among chemotherapy patients due to their vulnerability and frequent comorbidity. Notably, we confirmed in this study that exercise is a tolerable non-pharmacological option during chemotherapy treatment. This is consistent with the findings of a recent meta-analysis, which reported the absence of any harmful effects of exercise on cancer patients undergoing systemic treatment [33]. Another meta-analysis concluded exercise safety and feasibility among colorectal cancer patients [35]. This indicates that chemotherapy survivors may receive exercise-based care without any concerns of harm to reduce the impact of cancer on quality of life (tertiary prevention) and, at the same time, decrease the cardiovascular and metabolic risk in this vulnerable population.

Strengths and limitations

Few previous meta-analyses have addressed exercise's efficacy and safety profile in preventing CIC [52,53,54]. However, they either focused on one specific oncology population (i.e., breast cancer patients), one particular chemotherapy agent, or on safety outcomes only. Whereas our study provided a more robust examination of both possible cardiac benefits and harms of training among all oncology chemotherapy survivors. We thoroughly analyzed the available evidence using data from 952 participants and generated important findings about the benefit of exercise on cardiac function and aerobic fitness among cancer survivors managed with chemotherapy.

Nevertheless, our study was prone to considerable limitations as the available data from RCT was incomplete, and the involved studies presented significant heterogeneities and risk of bias concerns that could distort the final interpretations. Additionally, we did not provide a subgroup analysis of different chemotherapeutic agents. Finally, we did not assess the contribution of exercise in altering the susceptibility to develop or exacerbate myocardial ischemia, peripheral artery disease, thromboembolic disease, and myocarditis/pericarditis among chemotherapy patients as the evaluation of these outcomes would require other biomarkers (troponin elevation, ECG changes, INR drop for patients taking anticoagulants, vascular imaging, etc.), which are not included in our study.

Implications and future perspectives

The cardiovascular complications of cytotoxic molecules regroup a large spectrum of diseases [2]. Our results demonstrated a very modest benefit of exercise on the cardiac function of patients receiving chemotherapeutic agents, thereby, its low suitability to counteract chemotherapy-induced heart dysfunction. However, there is a potential for other cardioprotective effects not evaluated in our study, such as anti-ischemic, anti-thrombotic, and anti-inflammatory effects on chemotherapy-exposed cardiovascular tissue. Hence, future research should analyze the preventive abilities of physical activity against CIC events that may not necessarily lead to altered cardiac function, such as ischemic heart disease, peripheral artery disease, venous thromboembolism, and inflammatory reactions of the heart layers (myocarditis, pericarditis). On the other hand, the findings of our study suggest that there is a need for effective pharmacological and non-pharmacological strategies to prevent the decline in cardiac function secondary to chemotherapy. The only medication approved by the United States Food and Drug Administration (FDA) and European Medicines Agency (EMA) to prevent anthracycline-related cardiomyopathy is dexrazoxane [55]. However, other treatments were also found to be effective in preventing CIC, such as statins, beta-blockers, angiotensin-converting enzyme inhibitors, and aldosterone receptor antagonists, particularly spironolactone [56]. Therefore, the effectiveness of such therapies should be further investigated, and once confirmed, they may be approved for clinical use. The good tolerability of physical training programs by chemotherapy patients should motivate more investigation about the other possible benefits of this type of care apart from enhancing cardiovascular function and preventing CIC.

Conclusion

Exercise has limited beneficial effects on cardiac function among chemotherapy patients, manifesting mainly as a relative boosting of aerobic fitness. Nevertheless, it is a safe and tolerable strategy that may hold other interesting advantages to cancer survivors worthy of investigation. Moreover, the fact that exercise did not show beneficial chronotropic effects (no changes in RHR and PHR) does not support the protective value of training programs against tachyarrhythmias associated with chemotherapeutic agents. The absence of improvement in CO, LVEF, SV, LVEDV, LVESV, GLS, and E/A ratio indicates the inefficacy of exercise in reducing chemotherapy-induced left ventricular dysfunction and heart failure. Despite the shown lack of proof of effectiveness, future studies should still search for any possible cardioprotective potentials of physical training during chemotherapy. Parallel to this, it is also necessary to identify pharmacological or non-pharmacological strategies other than exercise to antagonize the cardiovascular harms of different chemotherapeutic drugs effectively.