Association between urinary dickkopf-3, acute kidney injury, and subsequent loss of kidney function in patients undergoing cardiac surgery: an observational cohort study

Abstract

Background: Cardiac surgery is associated with a high risk of postoperative acute kidney injury (AKI) and subsequent loss of kidney function. We explored the clinical utility of urinary dickkopf-3 (DKK3), a renal tubular stress marker, for preoperative identification of patients at risk for AKI and subsequent kidney function loss. Methods: This observational cohort study included patients who had cardiac surgery in a derivation cohort and those who had cardiac surgery in a validation cohort (RenalRIP trial). The study comprised consecutive patients who had elective cardiac surgery at the Saarland University Medical Centre (Homburg, Germany; derivation cohort) and those undergoing elective cardiac surgery (selected on the basis of a Cleveland Clinical Foundation score of 6 or higher) who were enrolled in the prospective RenalRIP multicentre trial (validation cohort) and who were randomly assigned to remote ischaemic preconditioning or a sham procedure. The association between the ratio of preoperative urinary concentrations of DKK3 to creatinine (DKK3:creatinine) and postoperative AKI, defined according to the Kidney Disease Improving Global Outcomes criteria, and subsequent kidney function loss, as determined by estimated glomerular filtration rate, was assessed. Findings: In the 733 patient in the derivation cohort, urinary concentrations of DKK3 to creatinine that were higher than 471 pg/mg were associated with significantly increased risk for AKI (odds ratio [OR] 1·65, 95% CI 1·10–2·47, p=0·015), independent of baseline kidney function. Compared with clinical and other laboratory measurements, urinary concentrations of DKK3:creatinine significantly improved AKI prediction (net reclassification improvement 0·32, 95% CI 0·23–0·42, p<0·0001). High urinary DKK3:creatinine concentrations were independently associated with significantly lower kidney function at hospital discharge and after a median follow-up of 820 days (IQR 733–910). In the RenalRIP trial, preoperative urinary DKK3:creatinine concentrations higher than 471 pg/mg were associated with a significantly higher risk for AKI (OR 1·94, 95% CI 1·08–3·47, p=0·026), persistent renal dysfunction (OR 6·67, 1·67–26·61, p=0·0072), and dialysis dependency (OR 13·57, 1·50–122·77, p=0·020) after 90 days compared with DKK3:creatinine concentrations of 471 pg/mg or less. Urinary DKK3:creatinine concentrations higher than 471 pg/mg were associated with significantly higher risk for AKI (OR 2·79, 95% CI 1·45–5·37) and persistent renal dysfunction (OR 3·82, 1·32–11·05) only in patients having a sham procedure, but not remote ischaemic preconditioning (AKI OR 1·35, 0·76–2·39 and persistent renal dysfunction OR 1·05, 0·12–9·45). Interpretation: Preoperative urinary DKK3 is an independent predictor for postoperative AKI and for subsequent loss of kidney function. Urinary DKK3 might aid in the identification of patients in whom preventive treatment strategies are effective.