Acute kidney injury and mortality 1 year after major non-cardiac surgery

Abstract

Abstract Background Even mild and transient acute kidney injury (AKI), defined by increases in serum creatinine level, has been associated with adverse outcomes after major surgery. However, characteristic decreases in creatinine concentration during major illness could confound accurate assessment of postoperative AKI. Methods In a single-hospital, retrospective cohort study of non-cardiac surgery, the association between postoperative AKI, defined using the Kidney Disease: Improving Global Outcomes criteria, and 1-year survival was modelled using a multivariable Cox proportional hazards analysis. Factors associated with development of AKI were examined by means of multivariable logistic regression. Temporal changes in serum creatinine during and after the surgical admission in patients with and without AKI were compared. Results Some 1869 patients were included in the study, of whom 128 (6·8 per cent) sustained AKI (101 stage 1, 27 stage 2–3). Seventeen of the 128 patients with AKI (13·3 per cent) died in hospital compared with 16 of 1741 (0·9 per cent) without AKI (P < 0·001). By 1 year, 34 patients with AKI (26·6 per cent) had died compared with 106 (6·1 per cent) without AKI (P < 0·001). Over the 8–365 days after surgery, AKI was associated with an adjusted hazard ratio for death of 2·96 (95 per cent c.i. 1·86 to 4·71; P < 0·001). Among hospital survivors without AKI, the creatinine level fell consistently (median difference at discharge versus baseline –7 (i.q.r. –15 to 0) µmol/l), but not in those with AKI (0 (–16 to 26) µmol/l) (P < 0·001). Conclusion Although the majority of postoperative AKI was mild, there was a strong association with risk of death in the year after surgery. Underlying decreases in serum creatinine concentration after major surgery could lead to underestimation of AKI severity and overestimation of recovery.