Abstract Background The understanding of acute kidney injury (AKI) has evolved, but it remains uncertain whether outcomes have improved over time. This study aimed to describe temporal changes in outcomes among intensive care unit (ICU) patients with AKI in Denmark from 2010 to 2024. Methods This cohort study included adult patients with a first-time ICU admission, and AKI identified using creatinine changes in the seven days following ICU admission. The risks of chronic kidney disease (CKD), kidney failure (KF), and death were assessed one year after ICU admission across three periods: 2010–2014, 2015–2019, and 2020–2024. CKD was defined as an outpatient estimated glomerular filtration rate (eGFR) below 60 ml/min/1.73 m2 for more than 90 days. Similarly, KF was defined as an eGFR below 15 ml/min/1.73 m2, a hospital diagnosis, kidney transplantation, or initiation of chronic kidney replacement therapy. Follow-up was divided into seven to 89 days and from 90 to 365 days, and death was considered a competing risk for CKD and KF. Results A total of 48,529 patients with AKI were identified. The seven to 89-day mortality was 23.4% (95% confidence interval (CI): 22.5%—24.3%) in 2010–2014, compared to 22.3% (95% CI: 21.7%—22.9%) in 2015–2019, and 20.7% (95% CI: 20.2%—21.2%) in 2020–2024. The 90- to 365-day mortality was 10.3% (95% CI: 9.5%—11.0%) in 2010–2014, 10.2% (95% CI: 9.7%—10.7%) in 2015–2019, and 9.5% (95% CI: 9.1%—10.0%) in 2020–2024. The 90- to 365-day risk of CKD was 21.0% (95% CI: 19.9%—22.1%) in 2010–2014, compared to 19.8% (95% CI: 19.1%—20.6%) in 2015–2019 and 17.3% (95% CI: 16.7%—17.9%) in 2020–2024. The 90- to 365-day risk of KF was 2.2% (95% CI: 1.9%—2.6%) in 2010–2014, 2.1% (95% CI: 1.9%—2.3%) in 2015–2019, and 2.0% (95% CI: 1.8%-2.3%) in 2020–2024. Conclusions In this cohort of ICU patients with AKI, mortality and risk of CKD decreased over time, while the risk of KF remained stable. These improvements are encouraging and may reflect changes in ICU management or post-AKI care, although the observational design precludes inference about underlying mechanisms .
Christiansen et al. (2026) studied this question.