Abstract
OBJECTIVES: To analyse the trajectory of sleep quality changes and identify influencing factors among patients undergoing maintenance haemodialysis (MHD). DESIGN: Longitudinal study design. SETTING: The study was conducted in the nephrology departments of two tertiary hospitals in Urumqi, Xinjiang, China. PARTICIPANTS: This study employed convenience sampling and completed follow-up assessments for 282 patients with MHD between December 2024 and August 2025. OUTCOME MEASURES: Data collection used a general information questionnaire, the Pittsburgh Sleep Quality Index and the Self-Rating Depression Scale. Sleep quality assessment timepoints included baseline (T1), 3 months (T2) and 6 months (T3). Latent class growth models were used to identify heterogeneous trajectories of sleep quality. Univariable and multivariable logistic regression analyses were used to determine independent predictors of sleep disorders trajectory categories. RESULTS: Among 282 MHD patients, latent class growth modelling identified four distinct sleep trajectories: 'High-Slightly Increasing' (C1, 24.5%), 'Low-Slightly Increasing' (C2, 29.4%), 'High-Declining' (C3, 27.7%) and 'Moderate-Increasing' (C4, 18.4%). Multivariable analysis showed that, compared with C2, baseline depression significantly increased the odds of belonging to C1 (OR=8.53, p<0.001), C3 (OR=4.65, p<0.001) and C4 (OR=2.71, p=0.012). Similarly, pruritus (OR=2.46, p=0.019) and elevated C reactive protein (OR=1.03, p=0.015) were specific predictors for the C1 relative to C2. CONCLUSIONS: This study reveals four heterogeneous sleep trajectories in MHD patients, underscoring a dynamic view of sleep quality. Depression is an overarching risk factor for unfavourable trajectories, while pruritus and inflammation specifically predict persistent poor sleep. Early screening and targeted interventions against these factors are crucial to improve sleep quality in MHD care.