Association of ICU size and annual case volume of renal replacement therapy patients with mortality

Background We aimed to reveal whether the size of an intensive care unit (ICU) or its annual case volume of patients treated with renal replacement therapy (RRT) for acute kidney injury (AKI) is associated with hospital mortality. Methods This was a retrospective cohort study in the Finnish Intensive Care Consortium (FICC) database in 20072008. We divided the 23 FICC-member ICUs first into small or large according to ICU size, and second into low, medium, or high-volume tertiles according to annual case volume of patients with RRT. We compared crude hospital mortality, Simplified Acute Physiology Score (SAPS) II-, and case-mix-adjusted hospital mortality in small vs. large ICUs and in low- or medium-volume vs. high-volume ICUs. Results The median (interquartile range) annual case volume of patients with RRT for AKI per one ICU was 25 (1945). Patients in small or low-volume ICUs were older and less severely ill. Crude and SAPS II -adjusted hospital mortality rates were significantly higher in small ICUs but not significantly different in case volume tertiles. After adjusting for age, severity of illness, intensity of care, propensity to receive RRT, and day of RRT initiation, treatment in low or medium volume ICUs was associated with an increased risk for hospital mortality. Conclusions Crude and adjusted hospital mortality rates of patients treated with RRT for AKI were higher in small ICUs. Patients treated in high-volume ICUs had a decreased adjusted risk for hospital mortality compared to those in low-or medium volume ICUs.