Association between trajectory of triglyceride-glucose index and all-cause mortality in critically ill patients with atrial fibrillation: a retrospective cohort study.

Abstract

Introduction: Previous evidence showed that triglyceride-glucose (TyG) index is strongly associated with poor prognosis in atrial fibrillation (AF) in the general population. In critically ill patients, physiological stress may cause rapid fluctuation in the TyG index, making single measurements insufficient for prognosis assessment. Furthermore, the impact of TyG index trajectories on outcomes in critically ill patients with atrial fibrillation has not yet been well elucidated. Therefore, our study aimed to assess the association between TyG index trajectories in patients with AF in intensive care unit (ICU) and all-cause mortality at 30-day, 90-day, 180-day and 365-day follow-up.

Methods: We used data from Medical Information Mart for Intensive Care (MIMIC)-IV database. Patients diagnosed with AF in ICU were enrolled. We applied group-based trajectory modeling to identify distinct TyG index trajectories, selecting the optimal model based on the Bayesian information criterion (BIC), Akaike information criterion (AIC), average posterior probability (AvePP), and clinical interpretability. Kaplan-Meier survival curve was used to compare the mortality in AF patients with different TyG index trajectories. Hazard ratios (HRs) were calculated to elucidate the association between trajectories and prognosis in Cox proportional hazard models. Restricted cubic splines (RCS) were used to assess the relationship between TyG index and outcomes.

Results: A total of 1,108 AF patients were enrolled. Four TyG index trajectories were identified including: (1) traj1 group (TyG index stable at low level), (2) traj2 group (TyG index slowly ascending at moderate level), (3) traj3 group (TyG index ascending then descending at moderate high level) and (4) traj4 group (TyG index fluctuate at high level). The Traj4 group demonstrated significantly higher mortality rates at all time points (30-day, 90-day, 180-day and 365-day) compared to other trajectory groups. In addition, Cox proportional hazard models indicated that patients in traj4 group had higher risk of mortality compared to those in traj1 group at 30-day (HR = 1.71, 95% confidence interval [CI], 1.14-2.56), 90-day (HR = 1.67, 95% CI, 1.17-2.39), 180-day (HR = 1.44, 95% CI, 1.03-2.06) and 365-day (HR = 1.44, 95% CI, 1.04-1.98). Meanwhile, the RCS indicated a linear association between TyG index and all-cause mortality.

Conclusion: In critically ill patients with AF, TyG index trajectories were significantly associated with 30-day, 90-day, 180-day and 365-day all-cause mortality. This suggested that TyG index trajectories could serve as a robust indicator for risk stratification and prognosis assessment in ICU patients with AF.