Abstract

Background

Cardiovascular disease (CVD) has become the leading cause of competitive mortality in female breast cancer (BC). Regular aerobic exercise (AE) has been widely accepted as an effective intervention to reduce cardiovascular risk in a variety of different clinical conditions. This study is aimed at evaluating the efficacy and safety of AE on cardiovascular risk factors in female BC and assessing the quality of the synthesized evidence.

Methods

We searched five English databases (Cochrane Library, PubMed, Embase, Scopus, and Web of Science) from inception to January 2023. Randomized controlled trials (RCTs) and cohort trials studying the effects of AE intervention on cardiovascular disease risk in female breast cancer were included. We used Stata 16 for data synthesis, Risk of Bias 2, and the Newcastle–Ottawa Scale for methodological quality evaluation and assessed the certainty of the synthesized evidence in the Grading of Recommendations, Assessment, Development and Evaluations (GRADE) approach.

Results

Forty RCTs and 6 cohort trials involving 44,877 BC patients showed AE reduced the incidence of CVD events by 29.4% [risk ratio (RR) = 0.706, 95% confidence interval (CI) (0.659, 0.757), low certainty] and coronary artery disease events by 36% [RR = 0.640, 95% CI (0.561, 0.729), low certainty]. AE improved LVEF, and reduced weight and hip circumference. The subgroup analysis results showed that nonlinear AE increased VO₂max by 5.354 ml·kg·min^−1 [mean difference (MD) = 5.354, 95% CI (2.645, 8.062), very low certainty] and reduced fat mass by 4.256 kg [MD = 4.256, 95% CI (-3.839, -0.094), very low certainty]. While linear AE reduced low-density lipoprotein cholesterol (LDL-C) by 8.534 mg/dL [MD = -8.534, 95% CI (-15.511, -1.557), low certainty]. The sensitivity analysis results showed that each trial did not affect the impact index of the highly heterogeneous outcomes.

Conclusions

Our study indicates that AE has a positive effect in reducing cardiovascular risk factors. The individualization principle of AE deserves more attention in the future. This will provide new ideas to reduce CVD events and improve the quality of life in female BC patients. However, further research on AE in female BC should take into account long-term and well-designed administration to draw definitive conclusions.

Introduction

According to the latest global cancer statistics, female breast cancer (BC) has become the most prevalent cancer, with an estimated 2.3 million new cases, accounting for 11.7% of all cancer cases [1]. With the development and advancement of medical technology, the use of immune and endocrine therapy has increased the 5-year survival rate of BC by 10%- 20% [2, 3]. However, it is accompanied by a significantly higher risk of cardiovascular disease (CVD) with approximately 9.4% [4] compared with people who did not have BC [5, 6]. Deaths due to CVD accounted for 16.3% of deaths in BC patients [7] and all-cause mortality increased 3.8-fold in breast cancer patients who developed CVD compared to breast cancer patients who did not develop CVD [8]. CVD has become the leading cause of competitive mortality in BC women [9], due to the cardiotoxic effects of BC treatments (anthracycline chemotherapy, radiotherapy, and biotherapy) and overlapping risk factors for breast cancer and CVD (lack of exercise, obesity) [10]. Therefore, strategies to reduce the risk of CVD should ideally target both treatment- and patient-related factors [11].

Regular aerobic exercise (AE) has been widely accepted as an effective intervention to reduce cardiovascular risk in a variety of different clinical conditions [12]. It has been shown in preclinical studies that AE exerts cardioprotective effects in a model of chemotherapy-induced cardiotoxicity by being able to attenuate chemotherapeutic drug-induced oxidative stress and apoptosis [13, 14]. Lack of physical activity and poor cardiorespiratory fitness in BC patients is well documented [15]. To date, some systematic reviews and meta-analyses have been published about this topic [16,17,18,19].ÌýThese studies have mainly focused on chemotherapy-related cardiotoxicity and used VO₂max and LVEF as the tools of measurement. In addition, one systematic review [16] suggested that vigorous aerobic training performed with continuous or interval training mode should be considered. Furthermore, there are no reviews of body composition in previous literature. The linear aerobic exercise approach utilizes standard intensity, frequency, and duration parameters after an initial lead in period, with static increases in session duration [20]. The non-linear aerobic exercise approach considers the principles of exercise training in order to optimize the adaptations to the exercise stimulus. Sessions and weeks progress over the course of the prescription and vary between low intensity and moderate and high intensity training in order to target various physiological systems involved in the cardiopulmonary response to non-linear aerobic exercise approach [20]. Therefore this study is aimed to seek effective exercise strategies for reducing cardiovascular risk factors in BC patients through a systematic review and meta-analysis of AE intervention on cardiovascular risk factors in female BC. We also aimed to identify the appropriate exercise approach through analyzing the linear or non-linearÌýapproach to provide information on exercise prescriptions best suited for patients scheduled to receive BC therapy.

Methods

We performed this study in accordance with the guidance of the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement [21]. PRISMA checklist and abstract checklist were attached to Supplementary Materials 1 and 2. The systematic review was registered with PROSPERO (CRD42022342396).

Search strategy

We systematically searched five databases including PubMed, EMBASE, Cochrane Library, Scopus, and Web of Science from their inception up to January 31st, 2023. We referred to the retrieval method of "P + I" and searched with "MeSH terms and Emtree terms + Entry terms". The search terms included "Breast Neoplasms", "Breast Cancer", "Exercise", "Aerobic Exercise", "training exercise", "endurance exercise" and so on. The search strategy without limits to publication year, type, and status was described in detail in Supplementary Material 3.

Inclusion and exclusion criteria of the study

Types of studies

To collect additional data to explore the relevance of AE on cardiovascular events in BC patients, randomized controlled trials (RCTs) and cohort studies (CSs) were included. Trials that did not describe the randomization method in detail were considered non-randomized studies of interventions and were excluded. Animal studies were also excluded.

Types of participants

Studies that included participants of any age with histologically confirmed non-metastatic breast cancer. We did not restrict the type of treatment (chemotherapy, radiotherapy, or surgery). Studies that evaluated different types of cancer (breast, ovarian, rectal.) were only included if data from a breast cancer-only group were provided.

Types of interventions and control

Patients in the intervention group were treated with AE (continuous or interval; home-based or under professional supervision). Patients in the control group were asked to only perform their daily activities and not to begin any formal exercise training. Studies with the intervention involving diet or acute exercise were excluded.

Types of outcome measures

Primary outcomes

• (1) Incidence of CVD events. CVD events were defined as coronary artery disease (CAD), heart failure, valve abnormality, arrhythmia, stroke, pericardial complications, pulmonary hypertension, thromboembolic disease, or CVD death, occurring after study enrollment.

• (2) Changes in cardiopulmonary endurance. Cardiopulmonary endurance was measured by the maximum oxygen uptake (VO2 max). Training types (linear aerobic exercise or nonlinear aerobic exercise) were used as moderator variables.

Secondary outcomes

• (1) Cardiovascular risk score.

• (2) Lipid profile. Lipid profile includes low density lipoprotein cholesterol (LDL-C), high density lipoprotein cholesterol (HDL-C), triglyceride (TG), and total cholesterol (TC).

• (3) Body Composition. Body composition includes body mass index (BMI), body weight, waist circumference (WC), hip circumference (HC), lean mass, fat mass, body fat percentage (BF%), and waist-to-hip ratio.

• (4) Left ventricular ejection fraction (LVEF).

• (5) C-Reactive Protein (CRP).

Study selection and data extraction

Search results were imported to EndNote X9. The eligibility of retrieved studies was evaluated using the established inclusion and exclusion criteria. The contents of data extraction include first author, publication year, sample size, treatment received for breast cancer, basic characteristics (age, clinical type) of the included patients, duration and frequency of aerobic exercise, outcome measures. Studies without detailed information on outcome measures were excluded. The screening of studies and data extraction were performed independently by two reviewers (QJ and CM), and any differences were resolved by discussion or the decision of the third reviewer (JL).

Quality assessment

The Cochrane Collaboration's Risk of Bias 2 (RoB 2) tool was used to assess the methodological quality of the included RCT [22]. The following five domains were assessed: bias in the randomization process; bias from the intended intervention; missing outcome data; bias in outcome measures; and selective reporting of outcomes. Each domain was classified as low, some concern, and high risk of bias. The risk assessment for each entry using this tool finally resulted in a study quality assessment.

The quality of CS was evaluated using the corresponding Newcastle–Ottawa Scale (NOS, range 0–9) [23]. The study was scored based on 8 items in three categories: selection of participants, comparability between study groups, and measurement of exposure factors or results. Each entry is scored accordingly, with a total scale score of 9. A study scoring 7–9 is considered a high-quality study, 4–6 is a medium-quality study, and less than 4 is considered a low-quality study.

The quality assessment of each included study was performed independently by two reviewers (QJ and CM), and any disagreements were resolved by discussion or the decision of the third reviewer (BX).

Evidence synthesis

We used Stata 16 (Stata Corp LLC, College Station, TX, USA) for all statistical analyses. Continuous variable’s effect estimates were expressed as weighted mean differences (WMD) with a 95% confidence interval (CI). We did I² testing to assess the magnitude of the heterogeneity between studies, a value of I²â€‰> 50% was considered substantial heterogeneity [24]. The random-effects model was used when synthesizing data with high heterogeneity.

Additional analyses

Subgroup analysis

Subgroup analyses were performed, based on whether patients received linear or nonlinear aerobic exercise in the intervention group. Aerobic exercise was distinguished as a linear or non-linear exercise group based on the intensity, frequency and duration parameters of the exercise. If substantial heterogeneity existed, the source of heterogeneity was explored.

Sensitivity analysis

According to the Cochrane Handbook for Systematic Reviews of Interventions [24], I² values between 0 and 50% indicated that heterogeneity might not be important. Therefore, we eliminated the included studies with I²â€‰> 50% one by one and performed sensitivity analysis to identify possible highly influential studies. Studies were considered influential if the removal significantly changed the summary effect (i.e., change going from significant to non-significant).

Publication bias

Publication bias of the cumulative evidence among individual studies was evaluated using a graphical method of funnel plot and Egger’s test if at least 10 trials were included for the synthesized outcome [25]. If publication bias existed, the meta-trim-fill method was employed using Stata 16 to assess the potential impacts of publication bias of cumulative evidence.

Certainty of evidence

We assessed the certainty of the cumulative evidence using the Grading of Recommendations, Assessment, Development, and Evaluations (GRADE) system [26], where the certainty of the evidence was classified as high, moderate, low, or very low. Risk of bias, inconsistency, indirectness, imprecision, and publication bias were assessed. We present our findings in a summary of findings (SoF) table.

Results

Details of included trials

We obtained 12,678 records from the database search, and after the selection process, a total of 40 RCTs involving 2,129 participants were enrolled and 6 cohort trials involving 42,748 participants were enrolled. The selection process was summarized in the flowchart shown in Fig.Ìý1.

Fig.Ìý1

Flowchart of the selection process

General and intervention information of the cohort trials

The two reports included were a retrospective cohort trial report [27] and a prospective cohort trial report [28]. For the reporting of CVD events, the retrospective cohort trials reported stroke events and the prospective cohort trials reported heart failure events, and all trials reported the occurrence of CAD events. In the two cohort trial reports, the intervention groups were all set up respectively with AE intervention of different intensities in the low, medium, and high groups, and the control groups were all asked to only perform their daily activities and not to begin any formal exercise training. Therefore, we conducted a meta-analysis of 6 trials from 2 cohort reports based on exercise intensity to investigate the relevance of AE intervention on the occurrence of CVD and CAD events in BC patients. The characteristics of the 6 cohort trials were summarized in TableÌý1.

Table 1 Characteristics of included trails

General and intervention information of the RCTs

The three reports [39, 46, 50] contained 2 intervention groups respectively. Although both intervention groups were aerobic, the exercise protocols had slightly different. One group was a continuous AE with a continuous increase in exercise duration and intensity, which meets the definition of linear AE. While the other group was an intermittent AE with alternating intensity and duration, which meets the definition of nonlinear AE. So we divided each of the three reports into 2 trials for inclusion in the meta-analysis. The intervention group of one report [55] had the same intervention protocol for AE but was divided into Hispanic and non-Hispanic groups based on the race of the BC patients, so we extracted 2 trials from this study for inclusion in the meta-analysis. A total of 2,129 participants were randomized in the AE intervention group (n = 1,094) and the control group (n = 1,035). The characteristics of 40 RCTs were summarized in TableÌý1.

The current exercise oncology prescription design follows the design principles of linear and nonlinear exercise prescriptions [20]. The linear prescription design approach utilizes standard intensity, frequency, and duration parameters with static increases in session duration after an initial lead cycle. The nonlinear prescription design approach considered the principles of exercise training, tailored to the relative intensity of the individual, varying between low, moderate, and high intensity training. Fifteen RCTs of AE design followed linear design principles, and the other 25 trials followed nonlinear design principles.

Quality assessment of studies

Cohort studies

Since the 6 observational trials included were all cohort studies, the NOS corresponding to the cohort study was used to evaluate their quality. Three studies scored 9 and the other scored 8. Overall, the 6 cohort trials were high quality. Details of the NOS evaluation of the 6 cohort trials were provided in Supplementary Material 4.

Risk of bias of RCTs

We assessed the risk of bias in 40 included trials with the RoB 2 tool. Only one trial was assessed as ‘Low’ risk of bias, and other 39 trials were assessed as ‘Some concerns’. Most concerns were caused by the measurement of the outcomes, since the assessment of outcomes could be influenced by knowledge of interventions that patients received.ÌýThe summary of the risk of bias is shown in Fig.Ìý2.

Fig.Ìý2

Risk of bias of included RCTs

Primary outcomes

Effect of AE on CVD

Six cohort trials reported the occurrence of CVD and CAD events in BC patients. The meta-analysis showed that compared to the control group, AE reduced the incidence of CVD events by 29.4% [RR = 0.706, 95% CI (0.659, 0.757), P < 0.05] and CAD events by 36% [RR = 0.640, 95% CI (0.561, 0.729), P < 0.05] in female BC. There was high effect heterogeneity among the CVD trials (I²â€‰= 81.2%, p < 0.05), as shown in Fig.Ìý3. While there was low effect heterogeneity among the CAD trials (I²â€‰= 14.7%, p = 0.32), as shown in Fig.Ìý4.

Fig.Ìý3

Forest plot of cardiovascular disease (CVD) events in breast cancer (BC)

Fig.Ìý4

Forest plot of coronary artery disease (CAD) events in breast cancer (BC)

Effect of AE on cardiopulmonary endurance

Cardiopulmonary endurance is expressed as the VO₂max. Pooled data from 26 RCTs (1216 participants: cases = 625, controls = 591) showed that compared to the control group, AE elevated VO₂max by 3.86 ml·kg·min^−1 [95% CI (2.107, 5.613), p < 0.05)] in BC patients. There was high effect heterogeneity among the studies (I²â€‰= 90.9%, p = 0.000), as shown in Fig.Ìý5.

Fig.Ìý5

Forest plot of maximum oxygen uptake (VO₂max) in breast cancer (BC)

The heterogeneity between trials was large, and to elucidate the sources of heterogeneity, subgroup analyses were performed according to AE design principles: classification into linear AE and nonlinear AE. We used a random effects model for subgroup analysis in Fig.Ìý6. (1) Linear AE group: eleven RCTs [29, 30, 35, 37, 39, 40, 46, 48, 50, 51, 58] were included, and the results showed that linear AE treatment was significantly better than the control group in increasing VO₂max, with no statistically significant difference [WMD = 1.884, 95% CI (-0.154, 4.283), p = 0.124]. (2) Nonlinear AE group: fifteen RCTs [32, 34, 36, 38, 39, 43, 44, 46, 50, 52, 55,56,57, 60] were included and the results showed that nonlinear AE treatment was significantly better than the control group, which could significantly improve VO₂max by 5.354 ml·kg·min^−1, with a statistically significant difference [95% CI (2.645, 8.062), p = 0.000±Õ.

Fig.Ìý6

Forest plot of maximum oxygen uptake (VO₂max) subgroup analysis in breast cancer (BC)

Secondary outcomes

Effect of AE on the cardiovascular risk score

Cardiovascular scores were used in 3 RCTs to measure the effects of AE intervention: the Reynolds risk score (RRS) [31], the metabolic syndrome z-score [33], and the Framingham Risk Score (FRS) [35]. The results showed that AE significantly reduced cardiovascular risk scores compared with the control group. But the meta-analysis could not be performed because each score involved only one trial.

Effect of AE on body composition

Twelve RCTs [33, 38, 40, 42, 45, 50, 54, 56, 58, 61, 62] showed a consistent effect of AE on BMI, while the fixed-effect analysis showed no statistically significance. Eleven RCTs [33, 38, 42, 43, 45, 50, 56, 60,61,62,63] reported the effect of AE intervention on body weight in women with BC, and meta-analysis showed that AE significantly reduced body weight by 1.966 kg [95% CI (-3.839, -0.094), p = 0.040] compared with the control group. Four RCTs [33, 45, 61, 64] reported the effect of AE intervention on the HC in women with BC, and meta-analysis showed that AE reduced HC by 2.742 cm [95% CI (-4.278, -1.206), p = 0.000] compared with the control group. Four RCTs [45, 49, 59, 64] reported WC, four RCTs [54, 56, 58, 61] reported waist-to-hip ratio, and we performed a meta-analysis using a fixed-effects model, while the results showed the effect of AE intervention on the WC and waist-to-hip ratio was not statistically significant in BC patients. Seven RCTs [33, 40, 49, 50, 60, 65] reported the effect of AE intervention on fat mass, thirteen RCTs [30, 33, 43, 45, 49, 50, 54, 58, 60,61,62,63] reported BF%, and seven RCTs [33, 43, 45, 49, 50, 63] reported lean mass. Because of the high heterogeneity of the meta-analysis, we performed subgroup analyses based on exercise training principles (linear exercise or nonlinear exercise). The meta-analysis results showed that nonlinear AE significantly reduced fat mass by 4.256 kg [95% CI (-3.839, -0.094), p = 0.04], while the remaining results were not statistically significant. The results were shown in Supplementary Material 5.

Effect of AE on lipid profile

Five RCTs [47, 49, 56, 59, 62] reported the effect of AE intervention on triglyceride, and our meta-analysis using the fixed effect model showed no statistically significant results. Four RCTs [31, 47, 59, 62] reported the effect of AE on TC, and six RCTs reported HDL-C [31, 47, 49, 56, 59, 62]. Because of the high heterogeneity of their evidence, subgroup analyses based on principles of exercise training showed no statistical significance. Five RCTs reported LDL-C, two [49, 62] of which followed the linear design, and three [35, 47, 59] of which followed the nonlinear design. The results of the subgroup analysis we performed showed no statistically significant effect of nonlinear AE, while linear AE could reduce significantly the LDL-C by 8.534 mg/dL [95% CI (-15.511, -1.557), p = 0.017]. The results were shown in Supplementary Material 5.

Effect of AE on LVEF

Four RCTs [29, 32, 41, 51] reported the effect of AE intervention on LVEF, three in the linear [29, 41, 51] AE design, and one [32] in the nonlinear AE design. The meta-analysis showed that compared to the control group, AE significantly increased LVEF by 7.081, [95% CI (1.891, 12.272), p = 0.007]. The result was shown in Supplementary Material 5.

Effect of AE on C-reactive protein

Five [31, 41, 44, 47, 49] RCTs reported the effect of AE intervention on C-reactive protein, including two linear exercise design trials and four nonlinear exercise design trials. We performed subgroup analysis and found that the effect of either linear AE design or nonlinear AE design on c-reactive protein in women with BC was not statistically significant. The result was shown in Supplementary Material 5.

Publication bias

We assessed the publication bias with funnel plots and Egger’s test. The asymmetry of the funnel plots and Egger’s test suggest that VO₂max (P = 0.848) and weight (P = 0.14) had no significant risk of publication bias, while the BMI (P = 0.028) and BF% (P = 0.032) may have moderate publication bias. The funnel plots were shown in Supplementary Material 6.

We performed the meta-trim-fill analysis of the outcomes with publication bias. The combined effect value of BMI before the meta-trim-fill was -0.09 [95% CI (-0.249, 0.069)], and after the meta-trim-fill was -0.215 [95% CI (-0.36, -0.069)]. There was some difference in the combined effect value before and after the meta-trim-fill, suggesting that the BMI outcome may be subject to publication bias and that the result was less stable. The combined effect value of BF% before the meta-trim-fill was -0.146 [95% CI (-0.458, 0.166)], and after the meta-trim-fill was -0.264 [95% CI (-0.554, 0.026)]. The difference between the results before and after the meta-trim-fill was small, suggesting that the effect of publication bias on BF% was small. The two meta-trim-fill plots were shown in Supplementary Material 6.

Sensitivity analysis

To determine the impact of each trial on the effect index in highly heterogeneous outcomes, we used a sensitivity analysis in our meta-analyses. Finally, we did not observe the significant effects of any individual trial, as shown in Supplementary Material 7.

Certainty of evidence

We assessed 24 synthesized pieces of evidence with GRADE. A total of 10 of these outcomes were assessed as low certainty, and fourteen were very low certainty. The main reasons to downgrade the quality of evidence are the unsatisfactory risk of bias, the limited sample size of included trials, and high statistical heterogeneity. The summary of findings is shown in TableÌý2.

Table 2 Evidence profile

Discussion

Meta-analysis of trial results

In this study, we retrieved as many RCTs as possible and performed the meta-analyses to assess the effect of AE on cardiovascular risk factors in female BC and divided them into linear AE and nonlinear AE for subgroup analyses based on exercise training principles. AE reduced the risk of CAD events in female BC. In terms of body composition, AE reduced body weight and HC, whereas nonlinear AE reduced fat mass, but had no significant effect on other body components (BMI, lean mass, BF%, waist-to-hip ratio, and WC). The development or maintenance of lean mass (through AE) is an important component of weight management, can help prevent an age-associated decline in metabolic rate and gains in fat mass [63]. The reason for non-significant differences in percent body fat and could be attributed to factors such as dietary intake. As participants progressed through cancer treatment and during post treatment, the exercise may have caused an increase in appetite and subsequent increase in calorie intake. Therefore limiting nutritional intake (i.e. total calories) during exercise is essential for body composition management in BC patients. Other reasons for the lack of significant changes in body fat composition, waist circumference and waist-to-hip ratio may be attributed to the fact that weekly exercise time may not have been sufficient in the intervention group. The total minutes of exercise per week in the AE groups ranged from 90 to 180. This falls short of the report guidelines published in the America [66] of 180 to 300 min per week of beneficial body composition changes.

AE can improve cardiorespiratory fitness by increasing VO₂max, especially for nonlinear AE. AE had no significant effect on CRP and lipid profile (TC, TG, and HDL-C), but linear AE reduced LDL-C. A study directly comparing linear and nonlinear AE in cancer patients found that nonlinear AE resulted in more significant improvements in cardiorespiratory and physical function and a greater reduction than a linear exercise in nausea, vomiting, pain, and severity of adverse effects such as physical fatigue, which supports our findings [20]. We found that AE training significantly increased LVEF compared with controls, which is consistent with a previous systematic review of exercise interventions [67]. The four clinical trials involved included three linear AE training (two for continuous AE and one for intermittent AE) and only one clinical trial of nonlinear AE, and the effects of different exercise types on LVEF still need to be studied in large-scale clinical trials.

Strength and limitations

This is the first systematic evaluation and meta-analysis exploring the effects of AE intervention, particularly exercise design principles, on cardiovascular risk factors in female BC. Previous studies [16, 18, 68] have demonstrated the feasibility, safety, and overall effectiveness of aerobic training for patients with BC; however, there is no consensus regarding the most appropriate training mode to optimize training-induced adaptations. Most studies involving BC patients have prescribed aerobic training based on guidelines from the Clinical Society of Oncology of Australia (COSA) [69], the American College of Sport Medicine (ACSM) [66], and the American Cancer Society (ACS) [70]. Although it has been shown that AE can significantly improve patients' cardiorespiratory endurance, there are more training modes involving linear AE and fewer training modes involving nonlinear AE.

The present meta-analysis has some limitations. There is a high degree of heterogeneity in some results in this article and we believe that the sources of heterogeneity are as follows. According to our classification, one source of bias in the analysis was the method of assessment of cardiorespiratory fitness. Four studies [32, 34, 46, 51] indirectly assessed VO₂max, so there was a small probability of bias regarding the prescribed training intensity and the training-induced increase in VO₂max. This is because some of the trials included in some studies had small sample sizes and there was some risk of bias in several trials. Second, the number of long-term trials conducted is very limited and larger, longer-duration trials are needed to explore the effects of AE on long-term cardiovascular disease risk management. In addition, the risk of bias in the included trials needs to be considered, which seems to be mainly related to the difficulty of measuring the outcomes. Therefore, it is important to note that the bias assessment of these projects does not reflect the low quality of the study design, but rather expresses the unavoidable bias introduced by the lack of blinding of outcome assessors. Therefore, larger and better-blinded controlled trials should be conducted to properly address this issue.

Perspectives for future research

It is difficult to draw definitive conclusions because the quality of the pooled studies was not high. Nevertheless, our study suggests that AE reduces the risk of CVD and CAD events in female BC and has a positive effect in reducing body weight and HC and improving cardiorespiratory endurance. In terms of exercise design principles, nonlinear AE had a positive effect in reducing fat mass and improving cardiorespiratory endurance, but linear AE showed a positive effect in reducing LDL-C. AE may be an active strategy to effectively reduce cardiovascular risk factors in BC patients and provide options for the development of exercise rehabilitation programs for BC patients. In future, clinical trials (randomized, controlled, blank-controlled), the design and implementation process should take full account of the AE protocol design principles.

Conclusion

In summary, our findings indicated that AE increases the VO₂max in women with BC improving LVEF, and body composition (weight, HC). But there may be no difference regarding the lipid profile (TG, TC, HDL-C), and CRP. We also showed that nonlinear AE training is effective at increasing the VO₂max and reducing the fat mass in female BC. Therefore, high-quality and long-term RCTs are needed to provide data on the persistence of the effect and to strengthen the certainty of the evidence.