Kidney Week

Abstract: FR-PO082

Validation of Urine Clusterin and MCP1 in Predicting Drug-Induced AKI

Session Information

• AKI: Epidemiology, Risk Factors, Prevention - I
  November 03, 2023 | Location: Exhibit Hall, Pennsylvania Convention Center
  Abstract Time: 10:00 AM - 12:00 PM

Category: Pharmacology (PharmacoKinetics, -Dynamics, -Genomics)

• 2000 Pharmacology (PharmacoKinetics, -Dynamics, -Genomics)

Authors

• Da, Yi, National University Hospital, Singapore, Singapore, Singapore

Yi Da,

• Othman, Nor Islya Emma, National University Hospital, Singapore, Singapore, Singapore

Nor Islya Emma Othman,

• K Akalya, Akalya, National University Hospital, Singapore, Singapore, Singapore

Akalya K Akalya,

• Hong, Wei Zhen, National University Hospital, Singapore, Singapore, Singapore

Wei Zhen Hong,

• Koh, Liang Piu, National University Hospital, Singapore, Singapore, Singapore

Liang Piu Koh,

• Chua, Horng-Ruey, National University Hospital, Singapore, Singapore, Singapore

Horng-Ruey Chua,

Background

We have published that elevated urine clusterin, MCP1, β2MG, KIM1 and cystatin-C predict drug-induced AKI in patients receiving nephrotoxic drugs(vancomycin, aminoglycosides and calcineurin inhibitors)with a 1–3 day lead-time. We aim to validate the best-performing biomarkers for drug-induced AKI prediction in susceptible patients.

Methods

The reproducibility of the five proposed urine biomarkers to predict AKI in 13 drug-induced AKI patients and 13 controls was determined, to curate the best-performing three biomarkers using a multiplex assay. A prospective single-centre study of 137 patients receiving nephrotoxic drugs was subsequently conducted. Urine samples were collected 2–5 days before AKI onset by KDIGO criteria or before the end of nephrotoxic therapy in non-AKI patients. The primary analysis was the ability of the selected biomarkers to predict AKI with a 2-day lead time using individual ELISA.

Results

Urine clusterin, MCP1, β2MG yielded consistent AKI prediction with respective AUCs of 86%, 75%, 74%, and were superior to that of KIM1 and cystatin-C in the initial 26 patients. Of the validation cohort of 137 patients, 28% developed AKI. AKI and non-AKI patients had a similar mean age of 55 years, with AKI patients having a higher baseline eGFR than non-AKI patients (104 vs 98 mL/min/1.73m² respectively, p=0.01). Median levels of biomarkers were higher in eventual AKI cases vs non-AKI patients (p<0.0001 for clusterin and MCP1,p=0.03 for β2MG). Their AUC for AKI prediction was 73(64-82)% with clusterin, 77(68-86)% with MCP1, and 62(51-72)% with β2MG. We determined threshold levels of clusterin and/or MCP1 for optimal AKI prediction.

Conclusion

Urinary clusterin >150 ng/mL or MCP1 >200 pg/mL predicts drug-induced AKI, with best precision achieved by further elevations of both biomarkers.

Funding

• Government Support – Non-U.S.