Acute kidney injury (AKI) associated with high-dose vancomycin (VAN) therapy is a clinical concern, but no uniform diagnostic criteria exist. The AKI Network (AKIN) proposed new criteria to diagnose AKI based on abrupt changes in serum creatinine or urine output. We conducted a prospective observational study to determine the incidence and severity of AKI and associated outcomes using the AKIN criteria versus traditional definitions. Eligible patients (n = 227) were elderly (median, 70 years) and received VAN therapy for 8 days (median). AKI occurred in 43 patients (19%) using AKIN criteria at an onset of 6 days. AKI incidence was similar for patients with a trough level of ≥15 (24%; 17/72) versus <15 (17%; 26/155) μg/ml. Compared to non-AKI patients, more AKI patients resided in the intensive care unit (ICU) (47% [20/43] versus 27% [50/184]; P = 0.017), had a prior AKI episode (19% [8/43] versus 7% [5/184]; P = 0.001), and received vasopressor (28% [12/43] versus 14% [25/184]; P = 0.04) and/or nephrotoxins (84% [36/43] versus 67% [123/184]; P = 0.04). Seventeen of the AKI patients met traditional criteria, of whom more patients had stage 2 and 3 AKI (76% versus 8%; P = 0.0001), dosage adjustment (41% versus 15%) and renal consultation (35% versus 12%), prolonged length of stay after AKI (11 versus 7.5 days) and died (29% versus 12%) than those diagnosed by AKIN criteria (P value not significant). Use of AKIN criteria for AKI has the potential to improve care of VAN-treated patients by facilitating early detection of AKI and warrants confirmation in large prospective trials.