Search for a command to run...
Abstract Background Acute kidney injury (AKI) is common and associated with adverse long-term outcomes. Previously we have shown that a four biomarker model of soluble tumour necrosis factor receptor-1 and 2 (sTNFR1, sTNFR2), cystatin C and estimated Glomerular Filtration Rate (eGFR) measured 90 days after AKI performed well in predicting subsequent kidney disease progression. However, external validation in independent cohorts is essential to move these findings towards clinical application. Methods A prospective, observational cohort of adults with AKI within 72hrs of onset was assembled. Participants had study visits at time of AKI, then 30, 60 and 90 days later for biomarker sampling. Outcomes were assessed at 1 year, including Major Adverse Kidney Events (MAKE365, a composite of >25% decline in eGFR from baseline, kidney replacement therapy (KRT), or death) and kidney disease progression. Logistic regression models incorporating biomarker combinations were evaluated using area under the receiver operating characteristic curve (AUC). Results From 122 participants recruited at time of AKI, 89 survived and had biomarker measurements available from outpatient study visits. Of these, 35% developed MAKE365 and 30% had kidney disease progression at 1-year. The biomarker model (sTNFR1, sTNFR2, cystatin C, eGFR) measured at day 90 discriminated those with MAKE365 with an AUC of 0.79 (95% CI 0.68–0.91), and kidney disease progression with AUC 0.79 (95% CI 0.67 – 0.91). The biomarker model had comparable performance at earlier timepoints of day 30 and 60. Conclusions A biomarker panel comprising sTNFR1, sTNFR2, cystatin C, and eGFR reliably predicts adverse outcomes up to one year post-AKI. This provides external validation of findings from previous biomarker discovery studies, and shows how this biomarker combination could be used to identify patients at lowest risk. This may support biomarker-guided approaches for personalised post-AKI risk stratification and follow-up.