Avaliação da nefrotoxicidade de colistina e polimixina B no manejo de infecções por bactérias multirresistentes: uma revisão sistemática com meta-análise

Abstract

Polymyxins are polypeptide antibiotics, discovered in 1947, but with interrupted use in the 1980s due to many reports of adverse reactions, mainly neurological and renal reactions. With the increase in number of diseases caused by multidrug-resistant Gram-negative bacteria and the lack of new antibiotics capable of combating them, interest in this class of drugs has been resumed. Currently, only colistin and polymyxin B are used because of the high toxicity of the other components of this class. As a result, these antibiotics are being studied again according to current standards, in order to better understand their characteristics, especially on nephrotoxicity, wich is one of the major limitations of use. Objectives: This work aims to collect evidence on the prevalence of nephrotoxicity in patients treated with colistin and polymyxin B by conducting a systematic review and meta-analysis of observational studies and from this to verify which polymyxin is safer to be used in clinic. Methods: The search was carried out in the Pubmed, Scopus and DOAJ databases in September 2016. The elaboration of this work followed Cochrane's methodology and the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) recommendations. Two reviewers performed the search independently and a third reviewer was consulted in case of divergence. Longitudinal observational cohort studies which provided treatments with polymyxins and several patients who developed nephrotoxicity were included. The quality of the articles was evaluated using the NOS (New Castle Ottawa) instrument. The meta-analyzes were performed using CMA (Comprehensive Meta-Analysis) software, using the event rate as an effect measure, with a 95% confidence interval. It was also used the inverse-variance as a statistical method and the random effects model due to the design of the studies included. The Higgins inconsistency (I2) test was used to investigate heterogeneity, and also, the cumulative meta-analysis and sensitivity analyzes through the hypothetical removal of each study, meta-regression and meta-analysis of subgroups were performed. Results: The database search resulted in 489 articles. After applying the inclusion and exclusion criteria, 95 articles composed the systematic review and the meta-analysis of prevalence, with a total of 7,911 individuals evaluated for nephrotoxicity. Through meta-analyzes, it can be verified that the prevalence of nephrotoxicity was 26.7% [95% Confidence Interval (CI): 22.8-30.9%] for colistin, 29.8% (CI 23.8-36.7%) for polymyxin B; however, there was no significant difference (p = 0.720) among the groups treated, indicating that there is no superiority of one drug over another in terms of renal damage. But, it can be observed that nephrotoxicity was underestimated in earlier studies, in which only creatinine dosages were used to classify nephrotoxicity and in those whose renal damage was assessed as a secondary outcome. The lack of report standardization was the greatest limitation. Conclusions: The prevalence of nephrotoxicity was similar between colistin and polymyxin B, evidencing that both are nephrotoxic. In order to increase the quality of the nephrotoxicity reports, it is important to standardize these studies through the use of criteria to assess renal damage and the adequate report of this outcome, which allows a more accurate estimation of renal damage and, therefore, a greater control of this adverse reaction.