Abstract
BACKGROUND AND OBJECTIVES: Patients receiving in-center hemodialysis are at a high risk of coronavirus disease 2019 (COVID-19) infection. A reduction in hemodialysis frequency is one of the proposed measures for preventing COVID-19 infection. However, the predictors for determining an unsuccessful reduction in hemodialysis frequency are still lacking. MATERIALS AND METHODS: This retrospective observational study enrolled patients who were receiving long-term thrice-weekly hemodialysis at the Thammasat University Hospital in 2021 and who decreased their dialysis frequency to twice weekly during the COVID-19 outbreak. The outcomes were to determine the predictors and a prediction model of unsuccessful reduction in dialysis frequency at 4 weeks. Bootstrapping was performed for the purposes of internal validation. RESULTS: Of the 161 patients, 83 patients achieved a dialysis frequency reduction. Further, 33% and 82% of the patients failed to reduce their dialysis frequency at 4 and 8 weeks, respectively. The predictors for unsuccessful reduction were diabetes, congestive heart failure (CHF), pre-dialysis overhydration, set dry weight (DW), DW from bioelectrical impedance analysis, and the mean pre- and post-dialysis body weight. The final model including these predictors demonstrated an AUROC of 0.763 (95% CI 0.654-0.866) for the prediction of an unsuccessful reduction. CONCLUSIONS: The prediction score involving diabetes, CHF, pre-dialysis overhydration, DW difference, and net ultrafiltration demonstrated a good performance in predicting an unsuccessful reduction in hemodialysis frequency at 4 weeks.