Search

Abstract: Introduction: Acute kidney injury (AKI) is associated with significant morbidity and mortality. It affects approximately 4 to 7% of hospitalizations each year with a reported annual increase of 10% in incidence. Many studies have relied on administrative diagnostic codes to describe temporal trends in AKI, which have poor sensitivity in identifying AKI. Furthermore, few studies have shown demographic trends in AKI using serum creatinine-based definitions (KDIGO AKI). Objective: We aimed to identify trends in AKI using diagnostic codes and KDIGO AKI and associated severity, mortality, and demographic changes over a 10-year study period. We also aimed to identify change in recognition of AKI through diagnostic codes during the study period. Design, setting, and participants: The retrospective cohort was composed of adult patients admitted to hospital in Alberta, Canada from 2009 to 2018 using the Alberta Kidney Disease Network database. Exposure/Measure: AKI was assessed using validated KDIGO AKI definitions and AKI diagnostic codes. AKI associated in-hospital acute dialysis, in-hospital all-cause mortality, and 90-day post discharge all-cause mortality were assessed. We used generalized linear models with a Gaussian family to determine the absolute rates of AKI and mortality by year of incidence. We determined the sensitivity and specificity of AKI diagnostic codes using KDIGO AKI definitions as the standard of reference. Results: Between January 2009 and December 2018, we identified 348, 242 hospitalizations with an episode of AKI (12.3%). An increase in rates of AKI was seen using both AKI diagnostic codes and KDIGO AKI definitions with an unadjusted mean rate increase of 14.2/1000 hospitalizations [95% CI 12.7,15.6, p