Purpose: Studies on recovery from acute kidney injury (AKI) in ICU patients yield variable results. We assessed the impact of different recovery definitions, of different exclusion criteria, and of imputing missing baseline creatinine on AKI recovery in a heterogeneous ICU population.
Methods: Secondary analysis of the EPaNIC database. Recovery of kidney function in patients who developed AKI in ICU was assessed at hospital discharge. We studied recovery rates of different AKI stages with different definitions of recovery after inclusion or exclusion of non-survivors and in patients with or without chronic kidney disease (CKD). In addition, the impact of imputing missing baseline creatinine was investigated.
Results: A total of 1310 AKI patients were studied of which 977 were discharged alive from hospital. Rate of complete recovery (absence of KDIGO criteria) was markedly higher in survivors than in all AKI patients (79.5 vs 67.0%), especially for more severe forms of AKI. For patients with CKD, only the need for renal replacement therapy worsened kidney outcome as compared with no-CKD patients. Using stricter definitions of complete recovery significantly reduced its occurrence. New or worsening CKD occurred in 30% of AKI survivors. In no-CKD patients with available baseline creatinine, using an imputed baseline did not affect recovery. Patients with unavailable baseline creatinine were different from those with known baseline and revealed different recovery patterns.
Conclusion: These results indicate the need for rigorous description of AKI severity, the included population, definitions, and baseline creatinine handling in reports on AKI recovery.