BACKGROUND: The development of antibiotic resistance in Gram-negative bacteria is a serious global problem. Resistance limits the number of antibiotics which can successfully treat bacterial infections; consequently, clinicians are increasingly reintroducing polymyxins into clinical practice despite evidence of toxicity. This review summarizes current published literature reporting the efficacy/tolerability of polymyxins in patients with carbapenem-, multidrug-, or extensively-drug resistant Gram-negative infections.
METHODS: Eligibility criteria were developed using the PICOS framework. Searches of MEDLINE®, MEDLINE® In-process, Embase®, and Cochrane were conducted in November 2018. Eligible studies reported efficacy and/or tolerability outcomes for patients with resistant infections who were treated with polymyxins. Data were descriptively summarized.
RESULTS: 2,219 identified articles; 48 articles representing 46 unique studies met eligibility criteria (four randomized controlled trials [RCTs], 42 observational studies). All RCTs evaluated colistin; of the observational studies: colistin (n=35), polymyxin B (n=4), multiple polymyxins (n=3). Studies were heterogenous with respect to size, causative pathogens, and antibiotic susceptibility definition. One RCT reported infection-related mortality: 26.6% (monotherapy) vs. 21.2% (combination). Three RCTs reported Day 28-30 all-cause mortality rates of 24%-45%; there were no significant differences by formulation or monotherapy/combination therapy. All-cause mortality reported in 12 observational studies ranged from 11%-83%. Clinical response rate in one RCT: 21% (monotherapy) vs. 27% (combination), 10 observational studies: 20%-87% (median: 53%), one of which showed significantly lower rates with low-dose vs. high-dose colistin; clinical cure rates in one RCT: 67% (aerosolized) vs. 72% (intravenous), 11 observational studies: 18%-82% (median: 55%); microbiological cure in two RCTs: 45%-69%, 15 observational studies: 32%-88% (median: 67%). Nephrotoxicity (per RIFLE criteria) was measured in three RCTs, ranging from 15%-48%; a similar range was seen in three observational studies (7%-51%). Neurotoxicity was measured in 17 observational studies but was less common, with reported rates of 0%-15%.
CONCLUSIONS: Published literature demonstrates acceptable response rates but unacceptably high mortality rates. Clinically and economically impactful adverse events such as nephrotoxicity were also observed. Although a large degree of heterogeneity exists in the identified evidence base, available data suggest that polymyxins may be sub-optimal treatments for serious bacterial infections, with a need for alternative treatments with better efficacy and tolerability profiles.
