Long-term trajectories of cognitive function, fatigue, and quality of life in patients after pituitary adenoma surgery: A retrospective study.

Pituitary adenomas (PAs) are common intracranial tumors where surgery is the primary treatment. While effective, many patients face persistent cognitive deficits, chronic fatigue, and diminished health-related quality of life (HRQoL) long after resection. Current literature lacks dynamic, long-term perspectives on how these heterogeneous outcomes evolve and interact over time. This single-center retrospective cohort study (N = 295) tracked patients between June 2021 and December 2024 who underwent first-time PA resection. Longitudinal assessments from 6 months to 5 years postoperatively included the Montreal Cognitive Assessment (MoCA), Fatigue Severity Scale (FSS), and SF-36 survey. Latent class growth analysis (LCGA) was employed to identify distinct recovery trajectories. Multinomial logistic regression identified baseline predictors, and chi-square tests analyzed inter-trajectory associations. LCGA identified heterogeneous patterns: three classes for cognition (Stable-High, 54.9%; Moderate-Improving, 33.2%; Persistent-Deficient, 11.9%), three for fatigue, and distinct pathways for HRQoL. Key predictors for poor recovery trajectories included older age, lower education, Cushing's disease, and the need for postoperative hormone replacement. Notably, severe chronic fatigue was a potent predictor of poor mental HRQoL (aOR = 8.15, 95% CI: 3.50-18.99). A robust "symptom cluster" was identified, where cognitive impairment and severe fatigue consistently co-occurred with poor life satisfaction. Recovery after pituitary surgery follows diverse, interrelated long-term trajectories rather than a uniform path. Identifying this specific postoperative symptom cluster underscores the necessity of integrated, holistic management. Our findings facilitate early risk stratification and provide a foundation for developing personalized rehabilitation strategies targeting vulnerable patient populations.