The attributable mortality of new-onset acute kidney injury among critically ill patients: a propensity-matched analysis based on a multicentre prospective cohort study

Purpose This study aimed to evaluate the attributable mortality of new-onset acute kidney injury (AKI). Methods The data in the present study were derived from a multi-center, prospective cohort study in China that was performed at 18 Chinese ICUs. A propensity-matched analysis was performed between matched patients with and without AKI selected from all eligible patients to estimate the attributable mortality of new-onset AKI. Results A total of 2872 critically ill adult patients were eligible. The incidence of new-onset AKI was 29.1% (n = 837). After propensity score matching, 788 patients with AKI were matched 1:1 with 788 controls (patients without AKI). Thirty-day mortality was significantly higher among the patients with AKI than among their matched controls (25.5% versus 17.4%, p < 0.001). Subgroup analysis in terms of AKI classification showed that there was no significant difference (p = 0.509) in 30-day mortality between patients with stage 1 AKI and their matched controls. The attributable mortality values of stage 2 and stage 3 AKI were 12.4% [95% confidence interval (CI) 2.6–21.8%, p = 0.013] and 16.1% (95% CI 8.2–23.8%, p < 0.001), respectively. The attributable mortality of persistent AKI was 15.7% (95% CI 8.8–22.4%, p = 0.001), while no observable difference in 30-day mortality was identified between transient AKI patients and their matched non-AKI controls (p = 0.229). Conclusion The absolute excess 30-day mortality that is statistically attributable to new-onset AKI is substantial (8.1%) among general ICU patients. However, neither stage 1 AKI nor transient AKI increases 30-day mortality.