Supplementary Material for: Intensive care unit versus high-dependency care unit for patients with sepsis-associated acute kidney injury requiring continuous kidney replacement therapy: An observational nationwide database study

Background:
Sepsis is a life-threatening condition that often leads to severe acute kidney injury (AKI). Continuous kidney replacement therapy (CKRT) can be provided to critically ill patients in resource-rich settings such as intensive care units (ICUs). However, whether different medical staffing levels exert a significant effect on the outcomes of patients with sepsis-associated AKI treated with CKRT is unknown. We aimed to compare the outcomes of patients with sepsis-associated AKI who received CKRT upon admission to an ICU versus a high-dependency unit (HDU), where the staffing intensity is lower than that in ICUs.
Methods:
This retrospective cohort study used data retrieved from the Diagnosis Procedure Combination database and enrolled adult patients with sepsis admitted to an ICU or HDU between April 2018, and March 2023. We identified patients who initiated CKRT on admission and assigned them to the ICU or HDU group. We conducted matching weight analyses using propensity scores to compare the frequencies of in-hospital mortality and complications between the groups.
Results:
Of 81,934 patients with sepsis, CKRT for severe AKI was initiated in HDUs and ICUs for 1,136 and 4,622 patients, respectively. Patients in the ICU group had higher cardiovascular Sequential Organ Failure Assessment subscores, lower body mass index, worse consciousness levels, and a greater need for mechanical ventilation compared with those in the HDU group. After adjustment, in-hospital mortality did not differ significantly between the ICU and HDU groups (-1.9 percentage points, 95% confidence interval: -5.2 to 1.4). Moreover, significant differences were not observed in complications including gastrointestinal bleeding, infections, thromboembolism, or cerebrovascular events. These findings remained consistent across the subgroup analyses.
Conclusions:
In this large cohort study, in-hospital mortality or other clinical outcomes did not differ significantly between patients initially treated in ICUs and HDUs. Further investigations are warranted to understand the relevant factors that guide the triage of these patients.