Factors associated with fatigue in systemic lupus erythematosus: a cross-sectional structural equation modeling analysis of depressive symptoms as a potential mediator

Objectives This study examined physiological, psychological, and environmental factors associated with fatigue in patients with systemic lupus erythematosus (SLE) and assessed whether depressive symptoms statistically mediated the relationships between these factors and fatigue. Methods A convenience sample of 252 patients with SLE was recruited from the Department of Rheumatology and Immunology, Second Hospital of Hebei Medical University, between June 2023 and June 2024. Guided by the Theory of Unpleasant Symptoms (TOUS), candidate variables were classified into physiological (including disease activity on the SLEDAI-2K, anemia, biologic agent use, and sleep disturbance), psychological (anxiety and depressive symptoms), and environmental factors (including employment status and social support). Pain severity and other clinically recorded variables were entered as covariates. Validated measures included the Fatigue Severity Scale (FSS), Athens Insomnia Scale (AIS), Hospital Anxiety and Depression Scale (HADS), and Social Support Rating Scale (SSRS). Univariate analyses, Spearman correlation analysis, and multivariable linear regression were performed using SPSS 26.0. Structural equation modeling (SEM) and indirect-effect testing based on bootstrapping (2,000 resamples; percentile 95% confidence intervals) were performed in R (lavaan). Sensitivity analyses (robust estimation, a Bollen–Stine bootstrap, and covariate-adjusted and alternative mediator models) were also performed. Results Social support, depressive symptoms, sleep disturbance, pain severity, biologic agent use, and employment status were found to be independently related to fatigue. Moreover, social support and sleep disturbance each had a significant indirect relationship with fatigue via depressive symptoms. The indirect associations persisted after clinical covariates were specified as predictors of both depressive symptoms and fatigue. However, social support was not independently associated with the binary fatigue outcome (FSS ≥ 4; OR = 0.96 per SD, P = 0.837). Conclusions In this cross-sectional sample, social support, depressive symptoms, sleep disturbance, pain severity, biologic agent use, and employment status were associated with fatigue, and depressive symptoms showed a statistically significant indirect association with fatigue through pathways linking social support and sleep disturbance. Because the data are cross-sectional, these findings should be interpreted as exploratory indirect associations that are compatible with, but do not establish, a mediating mechanism, and as hypothesis-generating rather than evidence of causality or treatment effect. Systematic assessment of social support and depressive symptoms, together with dynamic evaluation of sleep disturbance and pain, may inform comprehensive fatigue assessment, while longitudinal and interventional studies are needed before firm clinical recommendations can be made.