Abstract: Objective To analyze the potential fatigue categories of patients with chronic heart failure (CHF), and explore their correlation with clinical indicators and prognosis, laying a theoretical foundation for the development of stratified management strategies. Methods This study enrolled 110 patients diagnosed with CHF at Taihe Hospital between January 2021 and October 2023. Latent Profile Analysis (LPA), a person-centered statistical approach, was employed to identify distinct latent classes of fatigue among the patient cohort. Comparative analyses of clinical data and prognostic outcomes were conducted across these fatigue classes, and multinomial logistic regression was utilized to ascertain the determinants of fatigue heterogeneity. Results The LPA results indicated that a three-class model (low, moderate, and high fatigue groups) provided the optimal fit, as demonstrated by the lowest values for the akaike information criterion (AIC; 6 977.902), bayesian information criterion (BIC; 7 058.563), and sample-adjusted BIC (aBIC; 6 988.798). The model exhibited high classification reliability, with an entropy value of 0.831, average posterior probabilities on the main diagonal exceeding 90%, and cross-classification probabilities of ≤7.5%. The cohort was classified into a low-fatigue group (n=27, 24.55%; mean total score=42.39±6.83), a moderate-fatigue group (n=49, 44.55%; mean total score=61.67±6.93), and a high-fatigue group (n=34, 30.91%; mean total score=77.91±6.56). Multinomial logistic regression analysis revealed that the insomnia severity index (ISI; OR=2.845) and the generalized anxiety disorder-7 (GAD-7; OR=1.892) were significant correlates of the high-fatigue profile. Furthermore, ejection fraction (OR=0.650、0.505) and the multidimensional scale of perceived social support (MSPSS; OR=0.785、0.209) were identified as significant factors associated with both the moderate- and high-fatigue profiles (all P<0.05). Over a one-year follow-up period, prognostic outcomes exhibited a significant gradient across the groups. The low-fatigue group demonstrated the lowest rates of all-cause mortality (3.70%), rehospitalization for heart failure (14.81%), and composite endpoint events (death or rehospitalization, 18.52%). These rates were intermediate in the moderate-fatigue group (10.20%, 24.49%, and 30.61%, respectively) and were substantially elevated in the high-fatigue group (26.47%, 50.00%, and 61.76%, respectively). The differences between the groups were statistically significant (all P<0.01). Conclusions The heterogeneity of fatigue in patients with CHF can be distinctly classified into three characteristic subgroups: low, moderate, and high. Psychological symptoms, such as insomnia and anxiety, alongside diminished levels of social support, are significantly correlated with a higher degree of fatigue. Moreover, the high-fatigue group is associated with the poorest prognosis. Individualized management should be implemented based on fatigue stratification to improve clinical outcomes.

Key words: Chronic heart failure, Fatigue, Latent profile analysis, Symptom heterogeneity, Prognostic evaluation, Psychological symptoms