Abstract: Objective  To explore the dynamic trajectories of post-dialysis fatigue and its influencing factors in maintenance hemodialysis (MHD) patients. Methods Using convenience sampling, 373 patients from a hemodialysis center were enrolled from January to June 2024. General information, the Brief Fatigue Inventory (BFI), and the Pittsburgh Sleep Quality Index (PSQI) were collected immediately after dialysis (T1), with BFI repeated at 1h (T2), 3h (T3), bedtime (T4), and the next day (T5). Latent class growth modeling (LCGM) identified fatigue trajectories, and multinomial logistic regression analyzed influencing factors. Results A 3-class LCGM model best fit the data: mild-rapid recovery (33.5%, C1), moderate-medium recovery (42.9%, C2), and severe-slow recovery (23.6%, C3). Logistic regression (reference: C1) showed C2 was significantly associated with sleep disorders (OR=0.141, P＜0.001), hypotension (OR=0.472, P=0.024), dialysis vintage (OR=1.278, P＜0.001), and hemoglobin (OR=0.939, P＜0.001); C3 was linked to female sex (OR=2.246, P=0.032), sleep disorders (OR=0.064, P＜0.001), hypotension (OR=0.275, P=0.002), dialysis vintage (OR=1.289, P=0.002), and hemoglobin (OR=0.939, P＜0.001). Conclusion Post-dialysis fatigue shows a declining trend but with population heterogeneity, suggesting personalized interventions targeting trajectory-specific factors.

Key words: Latent category growth model, Maintenance hemodialysis, Fatigue after dialysis, Longitudinal trajectory