Longitudinal changes in cardiorespiratory fitness and risk of depressive and anxiety disorders in a nationwide cohort of 7 million participants

While baseline cardiorespiratory fitness (CRF) is inversely associated with mental health disorders, the association of longitudinal CRF changes on depression and anxiety risk remains unclear in large-scale populations. This study investigated the association between changes in estimated CRF (eCRF) and subsequent risk of depressive and anxiety disorders in Korean adults. This nationwide cohort study analysed 7,007,488 Korean adults aged 19 to 64 years using National Health Insurance Service data (2011–2022). An eCRF prediction equation was developed using an external dataset and applied to estimate individual CRF from consecutive health examinations (2011–2012 and 2013–2014). eCRF percentage change was categorized into seven predefined groups. Participants were followed from January 1, 2015, through January 31, 2022, and those with prior depressive or anxiety disorders were excluded. Incident depressive (ICD-10: F32–F33) and anxiety (F40–F41) disorders were identified from claims data. Multivariable Cox proportional hazards regression models assessed associations between eCRF changes and incident depressive and anxiety disorders. Marginal structural models with inverse probability of treatment weighting addressed time-varying confounding. Among the participants, mean age was 47.7 ± 10.0 years and 3,011,020 (42.9%) were female. During follow-up, 452,464 depressive and 712,377 anxiety disorders occurred. Compared with maintained eCRF, > 5% decrease was associated with increased risk of depressive disorders (HR [hazard ratio] = 1.168, 95% CI [confidence interval] 1.155–1.182) and anxiety disorders (HR = 1.119, 95% CI 1.108–1.129). Conversely, ≥ 5% increase was associated with reduced risk of depressive disorder (HR = 0.928, 95% CI 0.912–0.945) and anxiety disorder (HR = 0.860, 95% CI 0.848–0.873). Declining CRF was associated with increased risk of depressive and anxiety disorders, while improving CRF was associated with lower risk. Associations remained robust in analyses accounting for time-varying confounding. These findings suggest CRF changes as a potential indicator of mental health risk.