Crit Care Med. 2021 Apr 8. doi: 10.1097/CCM.0000000000005008. Online ahead of print.
ABSTRACT
OBJECTIVES: Acute kidney injury, acute kidney injury severity, and acute kidney injury duration are associated with both short- and long-term outcomes. Despite recent definitions, only few studies assessed pattern of renal recovery and time-dependent competing risks are usually disregarded. Our objective was to describe pattern of acute kidney injury recovery, change of transition probability over time and their risk factors.
DESIGN: Monocenter retrospective cohort study. Acute kidney injury was defined according to Kidney Disease Improving Global Outcomes definition. Renal recovery was defined as normalization of both serum creatinine and urine output criteria. Competing risk analysis, time-inhomogeneous Markov model, and group-based trajectory modeling were performed.
SETTING: Monocenter study.
PATIENTS: Consecutive patients admitted in ICU from July 2018 to December 2018 were included.
INTERVENTIONS: None.
MEASUREMENTS AND MAIN RESULTS: Three-hundred fifty patients were included. Acute kidney injury occurred in 166 patients at ICU admission, including 64 patients (38.6%) classified as acute kidney disease according to Acute Disease Quality Initiative definition and 44 patients (26.5%) who could not be classified. Cumulative incidence of recovery was 25 % at day 2 (95% CI, 18-32%) and 35% at day 7 (95% CI, 28-42%). After adjustment, need for mechanical ventilation (subdistribution hazard ratio, 0.42; 95% CI, 0.23-0.74) and severity of the acute kidney injury (stage 3 vs stage 1 subdistribution hazard ratio, 0.11; 95% CI, 0.03-0.35) were associated with lack of recovery. Group-based trajectory modeling identified three clusters of temporal changes in this setting, associated with both acute kidney injury recovery and patients' outcomes.
CONCLUSIONS: In this study, we demonstrate Acute Disease Quality Initiative to allow recovery pattern classification in 75% of critically ill patients. Our study underlines the need to take into account competing risk factors when assessing recovery pattern in critically ill patients.
PMID:33826581 | DOI:10.1097/CCM.0000000000005008