729

Abstract

Introduction: Postoperative acute kidney injury (AKI), characterized by increases in routinely measured serum creatinine (sCr) levels, is not only one of the most common complications affecting up to 30% of surgical patients, but is also associated with a 2 to 10 fold increase in hospital mortality. We assessed the effect of AKI on increase in hospital costs. Hypothesis: We hypothesize that postoperative AKI may increase hospital costs significantly. Methods: We conducted a retrospective cohort study of 27,841 adult surgical patients with no previous history of chronic kidney disease that were admitted to Shands Hospital between 2000 and 2010. AKI was defined by consensus RIFLE (Risk, Injury, Failure, Loss, End-stage) criteria. We modeled hospital cost using multiple linear regression to see the effect of AKI on hospital costs adjusting for patient demographic and clinical data. LASSO method was used to select variables for the model. We checked diagnostic plots to see if assumptions of the model were satisfied. Results: Prevalence of RIFLE-AKI was 37% and in-hospital mortality rate was 8% for patients with AKI while it was 0.6% for patients without AKI. Median (25th-75th percentile) hospital cost was $11734 (8088,18024) for patients without AKI compared to $26570 (14243, 53536) for those with AKI, with higher costs for more severe cases of AKI ($20,739 for Risk, $33,982 for Injury, and 68,267 for Failure). In the presence of AKI, median hospital cost is doubled for survivors and is tripled for nonsurvivors. Multiple linear regression model showed that male gender, surgery admission (vs. medicine), emergency admission (vs. routine elective), private insurance type, ICU admission, longer length of stay, in-hospital mortality, number of comorbidities, AKI and interaction of AKI with mortality are significantly associated with higher hospital cost. According to our model, AKI increases hospital cost by $2,800 for survivors and by $36,420 for nonsurvivors. Conclusions: Hospital cost increases significantly for patients with AKI adjusting for demographic and clinical characteristics, especially for patients who decease in hospital.