Acute kidney injury during colistin therapy: a prospective study in patients with extensively-drug resistant Acinetobacter baumannii infections.

The study aimed to prospectively assess incidence and risk factors for colistin-associated nephrotoxicity. This is a secondary analysis of a multicentre, randomized clinical trial, comparing efficacy and safety of colistin versus the combination of colistin plus rifampicin in severe infections due to extensively drug-resistant (XDR) Acinetobacter baumannii. The primary end point was acute kidney injury (AKI) during colistin treatment, assessed using the AKI Network Criteria, and considering death as a competing risk. A total of 166 adult patients without baseline kidney disease on renal replacement therapy were studied. All had life-threatening infections due to colistin-susceptible XDR A. baumannii. Patients received colistin intravenously at the same initial dose (2 million international units (MIU) every 8 h) with predefined dose adjustments according to the actual renal function. Serum creatinine was measured at baseline and at days 4, 7, 11, 14 and 21 (or last day of therapy when discontinued earlier). Outcomes assessed were 'time to any kidney injury' (AKI stages 1-3) and 'time to severe kidney injury' (considering only AKI stages 2-3 as events). When evaluating overall mortality, AKI occurrence was modelled as a time-dependent variable. AKI was observed in 84 patients (50.6%, stage 1 in 40.4%), with an incidence rate of 5/100 person-days (95% CI 4-6.2). Risk estimates of AKI at 7 and 14 days were 30.6% and 58.8%. Age and previous chronic kidney disease were significantly associated with any AKI in multivariable analysis. Neither 'any' nor 'severe AKI' were associated with on-treatment mortality (p 0.32 and p 0.54, respectively). AKI occurs in one-third to one-half of colistin-treated patients and is more likely in elderly patients and in patients with kidney disease. As no impact of colistin-associated AKI on mortality was found, this adverse event should not represent a reason for withholding colistin therapy, whenever indicated.