RESUMEN
AIM: Cisplatin causes acute kidney injury (AKI) in approximately one third of patients. Serum creatinine and urinary output are poor markers of cisplatin-induced AKI. Metabolomics was utilized to identify predictive or early diagnostic biomarkers of cisplatin-induced AKI. METHODS: Thirty-one adult head and neck cancer patients receiving cisplatin (dose ≥70 mg/m2 ) were recruited for metabolomics analysis. Urine and serum samples were collected prior to cisplatin (pre), 24-48 h after cisplatin (24-48 h) and 5-14 days (post) after cisplatin. Based on serum creatinine concentrations measured at the post timepoint, 11/31 patients were classified with clinical AKI. Untargeted metabolomics was performed using liquid chromatography-mass spectrometry (LC-MS). RESULTS: Metabolic discrimination was observed between "AKI" patients and "no AKI" patients at all timepoints. Urinary glycine, hippuric acid sulfate, 3-hydroxydecanedioc acid and suberate were significantly different between AKI patients and no AKI patients prior to cisplatin infusion. Urinary glycine and hippuric acid sulfate were lower (-2.22-fold and -8.85-fold), whereas 3-hydroxydecanedioc acid and suberate were higher (3.62-fold and 1.91-fold) in AKI patients relative to no AKI patients. Several urine and serum metabolites were found to be altered 24-48 h following cisplatin infusion, particularly metabolites involved with mitochondrial energetics. CONCLUSIONS: We propose glycine, hippuric acid sulfate, 3-hydroxydecanedioc acid and suberate as predictive biomarkers of predisposition to cisplatin-induced AKI. Metabolites indicative of mitochondrial dysfunction may serve as early markers of subclinical AKI.