Sleep Quality and Its Independent Correlates in Patients with Diabetic Kidney Disease: A Retrospective Network.

Background: Poor sleep quality is relatively common among patients with diabetic kidney disease, yet its multidimensional biopsychosocial independent correlates have not yet been fully elucidated. Purpose: This study aims to explore the independent correlates of poor sleep quality in patients with diabetic kidney disease and to describe the network structure relationships between poor sleep quality and its associated factors. Methods: This retrospective cross-sectional observational study included 336 patients with diabetic kidney disease admitted to hospitals between 2024 and 2025. Twenty-seven biopsychosocial variables were extracted from electronic health records. Independent correlated variables were screened using LASSO regression and multivariate logistic regression. Network analysis was performed via 1000 bootstrap resamples to estimate expected influence (EI) and bridge expected influence (BEI). Results: Among patients with diabetic kidney disease, the prevalence of poor sleep quality was 40.2%. The LASSO method identified 19 candidate variables, whereas multivariate logistic regression retained 13 independent, significant variables. Caffeine intake, smoking history, restless legs syndrome, anxiety, peripheral neuropathy, and nocturia were positively correlated with poor sleep quality. In contrast, glucocorticoid use and elevated SOD showed a negative correlation with poor sleep quality. In the network analysis, the strongest direct connections associated with poor sleep quality were observed for caffeine intake (r=0.842), anxiety (r=0.629), restless legs syndrome (r=0.526), peripheral neuropathy (r=0.430), and nocturia (r=0.378). The EI value for caffeine intake was the highest (1.323), followed by glucocorticoid use (0.461); the bridging connectivity for caffeine intake (BEI=1.671) and glucocorticoid use (BEI=0.729) was relatively high. The Bootstrap analysis indicated that the network stability was acceptable (CS coefficient = 0.470). Conclusion: Among patients with DKD, poor sleep quality is associated with various biopsychosocial factors. Caffeine intake, anxiety, restless legs syndrome, peripheral neuropathy, and nocturia exhibit a strong direct association with poor sleep quality. These variables may serve as important candidate factors in future prospective studies; however, their causal relationships require further validation.