Search

What is known and objective Infections due to multidrug-resistant gram-negative bacteria ( MDR- GNB) are a significant burden to the healthcare system globally. Colistin is the drug of choice for MDR- GNB and recent studies recommend high doses. This study investigated the safety of low-dose colistin and the relationship of minimum inhibitory concentration ( MIC) of colistin with bacterial cure in the treatment for MDR- GNB infections. Methods Computerized dispensing records identified all patients who received colistin during January 2010 and December 2011. Patients who were aged < 12 years old, who received colistin for < 72 h or had moderate to severe renal impairment were excluded. Medical records of the remaining patients were reviewed for the necessary data to determine the bacterial cure and nephrotoxicity of colistin. Multivariate logistic regression analysis was used to determine the predictors of bacterial cure. Results A total of 125 evaluable patients received colistin during the study period. Ninety-four of 125 (75·2%) patients achieved bacterial cure. No statistically significant differences were observed between patients who achieved and failed to achieve bacterial cure with regards to age, gender, site of infection, mg/kg dose or duration of colistin use. The average MIC in the bacterial cure group was significantly lower than the MIC in the bacterial failure group ( P = 0·002). Similarly, 30-day mortality from the last dose of colistin was significantly lower in the bacterial cure group ( P = 0·002). Nephrotoxicity occurred in 12·8% of patients and was not associated with the dose of colistin or concomitant use of nephrotoxic medications. MIC of <1 μg/mL was the only significant independent predictor of bacterial cure in the multivariate logistic regression analysis ( P = 0·015), whereas infection caused by MDR Klebsiella pneumonia was an independent risk factor for bacterial failure ( P = 0·049). What is new and conclusion Low-dose colistin is an effective option in the treatment for infections caused by MDR- GNB with a low incidence of nephrotoxicity. Patients who achieved bacterial cure had significantly lower MIC values of colistin against MDR- GNB than those who failed to achieve it. Colistin dose should be based on the MIC data of a given patient or local antimicrobial sensitivity data to maximize its efficacy. [ABSTRACT FROM AUTHOR]