Cardiorespiratory fitness as a marker of brain aging: how does the heart train the brain?

This editorial refers to ‘Associations of cardiorespiratory fitness indicators with brain volumes and cognitive function in patients with coronary artery disease: findings from the heart–brain trial’, by A. Carlén et al. https://doi.org/10.1093/eurjpc/zwaf765. Coronary artery disease (CAD) is no longer viewed as a condition confined to the epicardial coronary circulation. Over the past two decades, a growing body of evidence has reframed CAD as a systemic vascular disorder with multisystem consequences, including adverse effects on brain structure and function. Epidemiological studies consistently demonstrate that patients with CAD exhibit an increased risk of cognitive impairment, accelerated cognitive decline, and dementia compared with age-matched individuals without cardiovascular disease.1,2 Neuroimaging studies further support this link, revealing reduced total brain volume, greater white matter disease burden, and features of accelerated brain aging in individuals with ischaemic heart disease.3 The mechanistic underpinnings of this heart-brain connection are complex and multifactorial. Shared risk factors such as aging, hypertension, diabetes, smoking, and dyslipidemia undoubtedly contribute, but growing attention has focused on vascular inflammation, endothelial dysfunction, impaired cerebral perfusion, autonomic imbalance, and reduced cardiorespiratory reserve as convergent pathways linking CAD to adverse brain outcomes. Within this framework, cardiorespiratory fitness (CRF) has emerged as a particularly compelling integrative biomarker. CRF reflects the coordinated function of the cardiovascular, pulmonary, and muscular systems and is one of the strongest predictors of morbidity and mortality across clinical populations, including patients with established CAD.4 In populations free of overt cardiovascular disease, higher CRF has been associated with better cognitive performance, preserved brain volumes, and reduced risk of dementia.5 Aerobic fitness has been linked to larger hippocampal volume, improved executive function, and even reversal of age-related brain atrophy in randomized exercise interventions.6 However, whether these associations extend to patients with established CAD, who may already exhibit structural and functional brain vulnerability, remains insufficiently explored. Moreover, prior studies have focused almost exclusively on maximal oxygen uptake (VO2peak) as the primary measure of CRF. While VO2peak is the gold standard, cardiopulmonary exercise testing (CPET) yields a rich array of additional maximal and submaximal indices, such as oxygen uptake efficiency slope (OUES), ventilatory anaerobic threshold (VAT), oxygen pulse, and heart rate recovery, each reflecting distinct aspects of cardiorespiratory and autonomic physiology.7 The relevance of these alternative CRF indicators to brain health, particularly in CAD, remains largely unexplored. Against this backdrop, in this issue of the European Journal of Preventive Cardiology, the Heart–Brain trial investigators8 present a timely and comprehensive cross-sectional analysis examining associations between multiple CRF indicators, brain volumes, brain age, and cognitive performance in patients with stable CAD. Their work substantially advances the field by broadening the conceptualization of CRF and positioning it as a potential biomarker of structural brain health in a high-risk cardiovascular population. Using data from 105 patients with stable CAD enrolled in the Heart-Brain randomized trial, Anna Carlén, Lucía Sánchez-Aranda, and colleagues8 performed an impressively detailed phenotyping that included CPET-derived CRF indices, multimodal brain MRI, and a comprehensive neuropsychological battery. Several important findings emerge. First, both maximal (VO2peak, time to exhaustion) and submaximal (OUES, oxygen pulse) indicators of CRF were positively associated with global brain volumes, particularly total brain volume and white matter volume.8 Notably, these associations persisted after adjustment for age, sex, education, BMI, APOE4 status, and intracranial volume, underscoring their robustness. Second, the hippocampus emerged as a particularly sensitive region. Associations between VO2peak or time to exhaustion and hippocampal volume were consistently stronger than those observed for global brain measures, reaching medium effect sizes by conventional benchmarks. Importantly, patients in the lowest VO2peak tertile exhibited a markedly higher prevalence of hippocampal atrophy compared with those in the middle and upper tertiles, reinforcing the clinical relevance of this finding. Third, several CRF indices (including OUES, oxygen pulse, and heart rate recovery) were inversely associated with brain-predicted age difference (brainPAD), indicating that higher cardiorespiratory efficiency and autonomic recovery were linked to a younger-appearing brain. This observation extends prior work on brain age estimation by demonstrating that physiological fitness parameters relate not only to absolute brain volumes but also to composite markers of brain aging. Fourth, associations between CRF and cognitive performance were fewer and weaker. Only VAT and OUES showed significant relationships with specific cognitive domains, namely working memory and executive function/attentional control, while maximal CRF indicators were not independently associated with cognition after adjustment.8 These findings support a model in which CRF is more closely linked to structural brain integrity and biological brain aging than to contemporaneous cognitive performance in patients with stable CAD. This dissociation has important conceptual and clinical implications, suggesting that structural brain vulnerability may precede detectable cognitive deficits in this population. The authors leverage a comprehensive CPET protocol and analyze a broad spectrum of CRF indicators. By moving beyond VO2peak, they capture multiple physiological dimensions, including ventilatory efficiency, stroke volume surrogates, and autonomic recovery. This approach enhances mechanistic insight and improves translational relevance, particularly for patients unable to achieve maximal effort. The neuroimaging methodology is rigorous and contemporary. The use of FreeSurfer for volumetric analyses, CentileBrain for hippocampal normative modelling, and BrainageR for brain age estimation provides robust and well-validated structural endpoints.9,10 Adjusting brainPAD analyses for image quality further strengthens confidence in the findings. The neuropsychological assessment is comprehensive and domain-specific, incorporating both traditional and NIH Toolbox measures. This allows for nuanced interpretation of cognitive domains rather than reliance on global screening instruments alone. Of note, the authors thoughtfully adjust for key confounders, including APOE4 genotype, education, and intracranial volume, and conduct sensitivity analyses accounting for beta-blocker use and vascular risk factors.8 Finally, the study addresses a clinically relevant and underexplored population. Patients with stable CAD represent a growing demographic in whom prevention of cognitive decline is increasingly recognized as a priority, yet data linking fitness, brain structure, and cognition in this group remain sparse. Despite the above-mentioned strengths, several limitations merit critical consideration. Foremost, the cross-sectional design precludes causal inference. It remains unclear whether higher CRF preserves brain structure or whether individuals with healthier brains are more capable of maintaining higher fitness levels. Reverse causality is a particular concern in studies linking physical performance to brain outcomes. Therefore, future investigations with longitudinal and interventional data are essential to establish directionality. Second, the sample size, while respectable for an imaging-intensive study, limits statistical power, particularly for detecting associations with cognitive outcomes and for exploring effect modification by sex. The low proportion of women reflects the epidemiology of CAD but nonetheless constrains generalizability and precludes definitive conclusions regarding sex-specific associations. Third, participants were relatively healthy, stable, and motivated CAD patients, as required by the trial’s inclusion criteria. As a result, findings may not extrapolate to older, frailer, or more comorbid CAD populations, in whom brain vulnerability may be greater and fitness levels lower. Fourth, multiple statistical comparisons raise the possibility of type I error. Although the authors appropriately emphasize consistency and biological plausibility, future studies with prespecified primary endpoints would strengthen inference. Finally, the interpretation of cognitive findings could be expanded. While the limited associations observed may reflect true dissociation between structure and function, alternative explanations include insufficient sensitivity of cognitive tests, cognitive reserve masking deficits, or lag effects whereby structural changes precede measurable cognitive decline. This study makes an important conceptual contribution by positioning CRF (measured through both maximal and submaximal CPET indices) as a biomarker of structural brain health in CAD. The demonstration that hippocampal volume and brain age are more strongly associated with fitness than cognitive performance suggests that brain MRI endpoints may serve as early indicators of vulnerability and therapeutic response. The prominence of the hippocampus aligns with experimental and clinical literature highlighting this region’s sensitivity to vascular health, inflammation, and neurotrophic signalling.11 The finding that low CRF is associated with a higher prevalence of hippocampal atrophy in CAD patients is particularly compelling and may have prognostic significance, given the hippocampus’s central role in memory and dementia risk. Equally important is the demonstration that submaximal indices such as OUES and time to exhaustion convey meaningful brain health information. This has practical implications for clinical testing, especially in patients who cannot achieve maximal exertion due to orthopaedic limitations, chronotropic incompetence, or beta-blocker therapy. From a clinical standpoint, all these observations reinforce the importance of CRF as a holistic marker of health in patients with CAD, extending beyond cardiovascular outcomes to encompass brain integrity (Figure 1). For clinicians, several practical messages emerge. Heart–Brain Crosstalk: Cardiorespiratory Fitness as a Determinant of Brain Aging First, assessment of CRF should be viewed not merely as a prognostic tool for mortality but as a potential indicator of brain health and aging. Where CPET is available, incorporating submaximal indices such as OUES or time to exhaustion may provide valuable information even when VO2peak is unobtainable. The association between low fitness and hippocampal atrophy underscores the potential role of exercise prescription as a neuroprotective strategy in CAD.12 While definitive evidence awaits interventional trials, clinicians should be encouraged by the biological plausibility and existing data supporting exercise as a low-risk, high-benefit intervention. The dissociation between brain structure and cognition highlights the importance of early prevention. By the time cognitive deficits become clinically apparent, structural damage may already be advanced. Targeting CRF earlier in the disease course may help preserve brain reserve. The findings of the Heart-Brain trial naturally point toward several avenues for future research. Longitudinal studies are needed to determine whether baseline CRF predicts trajectories of brain atrophy, brain aging, and cognitive decline in CAD. Integration of repeated CPET and MRI assessments would allow for mediation analyses testing whether exercise-induced improvements in CRF translate into preserved brain structure. Randomized exercise interventions represent a critical next step. While aerobic training has been shown to increase hippocampal volume in healthy older adults, similar trials in CAD populations are scarce. The ongoing Heart–Brain intervention phase may provide pivotal evidence in this regard. Future work should also expand neuroimaging endpoints to include white matter hyperintensities, cerebral perfusion, and vascular reactivity, which may be particularly relevant in CAD-related brain injury. Multimodal imaging could clarify whether CRF preferentially protects vascular vs. neurodegenerative pathways. Finally, greater attention to sex differences, cognitive reserve, and genetic modifiers such as APOE genotype will be essential to personalize preventive strategies. The Heart-Brain trial investigators provide compelling evidence that CRF, captured through both maximal and submaximal CPET indices, is closely linked to structural brain health and biological brain aging in patients with stable CAD. The preferential association with hippocampal volume and brain age, alongside weaker links to cognitive performance, suggests that CRF may serve as an early biomarker of brain vulnerability in this high-risk population. While causal inference awaits longitudinal and interventional data, this work meaningfully advances the heart–brain field and strengthens the rationale for targeting fitness as a modifiable determinant of brain health in cardiovascular disease. Gaetano Santulli (Conceptualization, Data curation, Supervision, Visualization, Writing—original draft, Writing—review & editing [lead]), Shivangi Pande (Writing—original draft [supporting]), and Pasquale Mone (Writing—original draft [supporting]) The Santulli Lab is currently supported in part by the National Institutes of Health (NIH): National Heart, Lung, and Blood Institute (NHLBI: R01-HL164772, R01-HL159062, R01-HL146691), National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK: R01-DK123259, R01-DK033823), and by the American Heart Association (AHA, 24IPA1268813), to G.S. S.P. is supported in part by the American Heart Association (AHA-26POST1563063).