Dynamic Patterns of Estimated Cardiorespiratory Fitness and Risk of Depressive Symptoms in Older Adults: A Longitudinal Analysis Using LTMLE

Technical Summary: Dynamic Patterns of Estimated Cardiorespiratory Fitness and Risk of Depressive Symptoms in Older Adults

Problem Statement
Late-life depression represents a significant global health burden, yet the causal impact of long-term cardiorespiratory fitness (CRF) dynamics on the onset of depressive symptoms remains under-explored. Conventional epidemiological models often fail to adequately account for time-varying confounding, where factors such as adiposity (BMI), smoking status, and metabolic biomarkers (HbA1c, SBP, cholesterol) both influence and are influenced by fitness levels over time. Furthermore, most existing literature relies on single-time-point assessments or traditional regression models that cannot capture the cumulative effects of cardiovascular health trajectories or the complex temporal sequences of exposure. There is a critical need to determine whether the "bottom-up" trajectory of CRF serves as a primary determinant for depressive symptoms and to understand if these benefits are static or require continuous maintenance.

Methodology
This study utilized a longitudinal analysis of data from the English Longitudinal Study of Ageing (ELSA), covering an eight-year period (2004–2012) across Waves 2, 4, and 6. The analytical sample comprised 7,109 community-dwelling adults aged 50 and older.

• Exposure Measurement: Cardiorespiratory fitness was estimated (e-CRF) using validated, sex-specific non-exercise equations based on anthropometric and physiological measures (age, BMI, waist circumference, resting heart rate, physical activity, and smoking status). e-CRF was dichotomized into high and low groups based on age- and sex-specific medians.
• Outcome Measurement: Depressive symptoms were assessed using the 8-item Center for Epidemiologic Studies Depression Scale (CES-D). A score of ≥3 was used as the clinical threshold for "caseness."
• Causal Inference Framework: The study employed Longitudinal Targeted Maximum Likelihood Estimation (LTMLE) integrated with Super Learning. This doubly robust, semi-parametric framework was selected to address time-varying confounding.
  • Time-Varying Confounders: The models adjusted for 22 time-varying covariates (L-nodes) including age, marital status, BMI, smoking history, systolic blood pressure (SBP), total cholesterol, and HbA1c.
  • Trajectory Analysis: Participants were categorized into dynamic exposure sequences (e.g., 1,1,1 for sustained high fitness; 1,0,0 for initial high fitness followed by decline; 0,0,0 for sustained low fitness).
  • Model Robustness: Seven hierarchical adjustment models were tested, ranging from basic sociodemographic adjustments to fully adjusted models incorporating all clinical and lifestyle markers. Sensitivity analyses included varying Super Learner algorithms (GLM, XGBoost, and ensembles) and comparing hybrid multiple imputation with complete-case analysis.

Key Contributions and Results
The study provides a rigorous causal evaluation of how dynamic e-CRF patterns influence depression risk, revealing several critical findings:

1. Sustained High Fitness as a Potent Protector: Participants maintaining high e-CRF across all three waves (trajectory 1,1,1) exhibited a robust and stable protective effect against depressive symptoms. In the fully adjusted model (Model 7), sustained high fitness was associated with a 63.3% reduction in risk (OR = 0.367; 95% CI: 0.295–0.457; P < 0.001) compared to the persistent low fitness group. This association remained remarkably consistent across all seven adjustment models, suggesting the neuroprotective benefits of fitness are largely independent of conventional metabolic and lifestyle factors.

2. The "Perishable" Nature of Fitness Benefits: The study identified a temporal dose-response relationship. While participants who maintained fitness through the study midpoint (trajectory 1,1,0) retained a significant 46.9% risk reduction (OR = 0.531; P < 0.001), the protective effect was entirely attenuated in the rapid-decline group (trajectory 1,0,0). In the fully adjusted model, the OR for the (1,0,0) group shifted to 0.869 (P > 0.05), indicating that early-life fitness does not provide a lasting "memory effect" against depression if followed by physical decline, weight gain, or smoking.

3. Independence from Metabolic Pathways: The stability of the odds ratios across models adjusting for metabolic biomarkers (SBP, HbA1c, cholesterol) suggests that the e-CRF–depression link is not merely a secondary effect of improved cardiovascular or metabolic health. This points toward direct neuroprotective mechanisms, such as increased BDNF expression or hippocampal volume modulation.

4. Therapeutic Window in Later Life: Trajectory analysis of the "Increasing" group (those transitioning from low to high fitness) showed a significant risk reduction (OR = 0.47), suggesting that the aging brain retains neuroplasticity and that fitness interventions initiated in later life can provide a "rescue" effect against depressive onset.

5. Subgroup Consistency: The protective effect of sustained high e-CRF was consistent across gender, smoking status, and age strata, with no significant statistical interactions found (Pinteraction > 0.05 for all), indicating a universal protective mechanism.

Significance and Claims
The authors claim this study is the first to explore the causal impact of e-CRF on mental health using a doubly robust method that explicitly accounts for time-varying confounding by BMI, smoking, and metabolic biomarkers.

• Methodological Advancement: By utilizing LTMLE and Super Learning, the study overcomes the "G-computation" challenge of time-varying confounding that limits previous fitness-depression research, offering a more reliable template for psychiatric epidemiology.
• Clinical Implications: The findings advocate for the integration of e-CRF monitoring into routine geriatric care. The authors suggest that clinicians should view a decline in cardiorespiratory fitness as an early warning sign for future depressive risk.
• Public Health Strategy: The study emphasizes that mental health benefits of fitness are dynamic and "perishable," requiring consistent maintenance rather than simple weight loss or sporadic activity. It supports multi-domain interventions that combine aerobic training with weight management and smoking cessation to ensure neuroprotective effects are not offset by metabolic decline.
• Optimistic Outlook: Contrary to "critical period" hypotheses, the results suggest that physical improvement in the short term provides a measurable "rescue" effect, offering a viable strategy for reducing the global burden of late-life depression through sustained physiological conditioning.