Exercise May Not Shield the Heart From Breast Cancer Drugs, Review Finds

For years, cardiologists and oncologists have urged women undergoing chemotherapy for breast cancer to stay active, hoping that structured exercise could protect the heart from the collateral damage of life-saving drugs. The idea is biologically plausible and clinically appealing: anthracyclines such as doxorubicin and targeted therapies like trastuzumab can quietly injure the heart muscle, and exercise is known to strengthen cardiovascular resilience in almost every other context. But a new systematic review and meta-analysis of randomized controlled trials, published in Supportive Care in Cancer, delivers a sobering reality check. After pooling the best available randomized evidence, researchers from the University of León and collaborating Spanish institutions found that the data simply do not confirm a consistent cardioprotective effect of exercise on the key measures of heart function used in cardio-oncology.

The research team, led by Raúl Cobreros-Mielgo together with Jesús Seco-Calvo and Diego Fernández-Lázaro, set out to answer a deceptively simple question: in women receiving anthracyclines and/or anti-HER2 therapy, does structured exercise, compared with usual care, preserve left ventricular ejection fraction (LVEF), global longitudinal strain (GLS), and cardiac biomarkers? These three measures form the backbone of modern cardio-oncology surveillance. LVEF, the percentage of blood the left ventricle pumps out with each contraction, is the classic workhorse metric of cardiac function. GLS, measured by speckle-tracking echocardiography, quantifies the subtle deformation of the heart muscle during systole and can detect subclinical injury before the ejection fraction falls. Cardiac biomarkers, particularly troponins and natriuretic peptides, signal myocardial stress and damage at the molecular level.

The methodological framework of the review was rigorous. The authors searched multiple databases and trial registries, following the PRISMA 2020 reporting standards, and assessed the risk of bias in each included trial using the revised Cochrane risk-of-bias tool for randomized trials, known as RoB 2. They then rated the certainty of the evidence with the GRADE approach, the international standard for grading how much confidence clinicians can place in a body of evidence. When trials were clinically and methodologically comparable, effects were pooled as net differences in change between the exercise and control groups, calculated as the change in the experimental group minus the change in the control group, using random-effects models. Crucially, the primary synthesis was restricted to the baseline-to-end-of-anthracycline-treatment window, a period of roughly four months, to ensure that the pooled estimates reflected a consistent phase of cardiotoxic exposure.

What they found was striking in its sparseness. After screening the literature, only seven randomized trials, reported across nine publications and encompassing a total of 513 women, met the inclusion criteria. Of these, quantitative pooling proved feasible in just two trials. Most of the included studies carried some concerns or a high risk of bias according to RoB 2, meaning that methodological weaknesses, such as problems with randomization, blinding, or missing outcome data, could plausibly have distorted their results. This is a common pattern in exercise oncology research, where blinding participants to physical activity is inherently impossible and sample sizes are often constrained by the difficulty of recruiting patients in the middle of cancer treatment.

The pooled results were essentially null. For LVEF, the mean difference between exercise and control groups was 0.07 percentage points, with a 95 percent confidence interval spanning from −1.38 to 1.52 and low heterogeneity between trials. For GLS, the mean difference was −0.10, with a confidence interval from −0.79 to 0.58 and no heterogeneity at all. In plain terms, the exercise groups and the control groups showed indistinguishable changes in both measures over the treatment window. But the authors are careful to emphasize what these numbers do and do not mean. The confidence intervals were wide, compatible with a meaningful benefit, with no effect, or even with harm. This is the classic statistical scenario in which absence of evidence cannot be mistaken for evidence of absence: the studies were too small and too few to rule out a clinically important cardioprotective effect, even though the point estimates hovered at zero.

Biomarker data proved even less tractable. The studies that measured troponins, natriuretic peptides, and other circulating markers of cardiac injury used such different assays, sampling schedules, and reporting conventions that the authors judged meta-analysis to be inappropriate. Instead, they described the findings narratively, and the heterogeneity prevented any unified conclusion about whether exercise blunts the molecular signature of chemotherapy-induced cardiac stress. Equally telling is what the trials never measured at all. No included study reported comparative clinical cardiovascular events, such as heart failure hospitalizations or arrhythmias, or survival outcomes. Nor did any trial formally assess heart failure with preserved ejection fraction, the diastolic dysfunction syndrome that is increasingly recognized as a late consequence of anthracycline exposure and that standard LVEF monitoring can miss entirely.

The implications reach well beyond this single review. International guidelines, including the 2022 European Society of Cardiology cardio-oncology guidelines and the American Heart Association’s scientific statement on cardio-oncology rehabilitation, currently recommend exercise training for patients receiving potentially cardiotoxic cancer therapies, largely on the basis of general cardiovascular benefit, fitness preservation, and quality-of-life gains. Those recommendations remain reasonable, and the new review does not contradict them for those broader outcomes. What it does challenge is the more specific and increasingly common claim that exercise reliably protects the myocardium itself during treatment. The authors point to prior systematic reviews that reached more optimistic conclusions and note that those analyses often pooled non-randomized studies, mixed treatment windows, and combined incomparable endpoints, inflating apparent certainty.

The mechanistic backdrop makes the question genuinely fascinating rather than merely academic. Anthracyclines damage cardiomyocytes through oxidative stress, mitochondrial dysfunction, and topoisomerase IIβ inhibition, while trastuzumab interferes with neuregulin-1 signaling that supports cardiac repair. Exercise, meanwhile, induces antioxidant defenses, improves endothelial function, and promotes mitochondrial biogenesis, all of which could theoretically buffer these insults. Preclinical work and small mechanistic studies, including trials of exercise preceding doxorubicin infusion, have hinted at protective signals. But the randomized clinical evidence reviewed here shows that translating those signals into measurable preservation of LVEF or GLS in real patients has not yet been demonstrated. Notably, the authors registered their protocol prospectively on PROSPERO and made their full analysis code, extraction sheets, and search strategies publicly available on GitHub and archived in Zenodo, an unusually transparent approach for this field.

Where does this leave patients and clinicians? The honest answer is in a state of productive uncertainty. Exercise during chemotherapy remains safe in appropriately screened patients and continues to deliver well-documented benefits for fatigue, physical function, and possibly treatment tolerance. But women and their care teams should not be told, on current evidence, that a training program will definitively protect their ejection fraction or strain measurements. The review’s authors call for larger, adequately powered randomized trials with standardized echocardiographic protocols, centralized strain analysis, harmonized biomarker sampling, and, critically, follow-up long enough to capture clinical events such as heart failure, not just surrogate imaging changes. Until such trials exist, the cardioprotective promise of exercise in cardio-oncology remains a hypothesis worth testing, not a proven prescription, and the 513 women studied so far represent only the first tentative steps toward an answer.

Subject of Research: Effects of structured exercise on cardiac outcomes during anthracycline or anti-HER2 breast cancer therapy

Article Title: Structured exercise and cardiac outcomes during anthracycline or anti-HER2 therapy in breast cancer: systematic review and meta-analysis of randomized trials

Article References: Cobreros-Mielgo, R., Seco-Calvo, J., & Fernández-Lázaro, D. (2026). Structured exercise and cardiac outcomes during anthracycline or anti-HER2 therapy in breast cancer: systematic review and meta-analysis of randomized trials. Supportive Care in Cancer, 34(10), Article 1049. https://doi.org/10.1007/s00520-026-11287-5

Image Credits: AI Generated

DOI: 10.1007/s00520-026-11287-5

Keywords: breast cancer, cardiotoxicity, exercise therapy, cardio-oncology, anthracyclines, trastuzumab, left ventricular ejection fraction, global longitudinal strain, systematic review, meta-analysis, heart failure, randomized controlled trials