Antimicrobial resistance (AMR) is estimated to cause 23,000 deaths annually in the United States and 700,000 worldwide. Existing surveillance systems track AMR within and across antibiotic categories, but do not estimate the burden of clinically meaningful co-resistance and its relationship to patient outcomes. A novel approach to tracking antimicrobial resistance in bacterial and fungal infections involves focusing on treatment-limiting co-resistance to all first-line agents, which can be described as difficult-to-treat resistance (DTR). Although strains susceptible to only one or two first-line agents also present therapeutic challenges for choosing empiric therapy, DTR signifies zero active first-line agents and an even higher level of resistance. Non-susceptibility to all first-line, high efficacy, low toxicity agents very often leads to discordant empirical regimens and subsequent reliance on less effective, and/or more toxic reserve agents. Therefore, DTR captures clinically impactful co-resistance that likely affects patient outcomes. To more fully describe trends and the impact of DTR, we used two large deidentified databases two separate systems (study 1: PremierTM from 173 hospitals; and study 2: Cerner Health Facts from 140 hospitals) with linked microbiologic, clinical, and demographic information. We evaluated >60 million patient encounters to identify those with gram-negative blood stream infections (GNBSI). DTR was defined as intermediate/resistant in vitro to all -lactam categories including carbapenems and fluoroquinolones; carbapenem-resistant CR, extended-spectrum cephalosporin-resistant ECR and fluoroquinolone-resistant FQR was defined using CDC surveillance definitions. The adjusted relative risk of mortality (aRR; Poisson regression) and burden of reserve agent use were compared across DTR and CDC-defined phenotypes susceptible to one first-line agent. Additionally, within study 2, we assessed the impact of adjustment for sepsis-related organ failure score (SOFA) on mortality. In study 1, 471 (1%) of 45,011 GNBSI episodes at 92/173 (53.2%) hospitals from 2009-2013 exhibited DTR, ranging from 0.04% for E coli to 18.4% for A. baumannii. Among DTR cases, 79% received parenteral aminoglycosides, tigecycline, or colistin/polymyxin-B; resistance to all aminoglycosides occurred in 33%. Predictors of DTR included urban healthcare and higher baseline illness. Crude mortality for DTR-GNBSIs was 43%; aRR was higher for DTR vs. CR (aRR=1.2, 95% CI=1.0-1.4; p=0.02), ECR (aRR=1.2, 1.1-1.4; p=0.001), or FQR (aRR=1.2, 1.0-1.4; p=0.008). Mortality aRR increased 20% per graded loss of active first-line categories from 3-5 to 1-2 to 0. In study 2, 25,448 unique encounters from 2009-2015 with DTR GNB were identified. Prevalence of DTR by species was similar to previously reported estimates. Crude mortality in patients with DTR GNB varied by taxon from 46% in Enterobacteriaceae to 59% in P. aeruginosa. Compared to CR patients with 1 active first line category available, DTR patients more frequently used colistin (9% vs. 5%), tigecycline (19% vs. 4%), and aminoglycosides (38% vs. 24%). aRR was higher for DTR vs. CR (aRR=1.29, 95% CI=1.02-1.64; p=0.04), ECR (aRR=1.45, 95% CI=1.17-1.78; p=0.0005), or FQR (aRR=1.52; 95% CI=1.23-1.88; p=0.0001). The aRR for mortality in DTR decreased by 7-21% upon adjustment using baseline SOFA score. Co-resistance to first-line agents necessitates use of reserve antibiotics and decreases survival in patients with GNB. Future assessments of clinical impact of resistance on outcome should include risk-adjustment using clinical and physiologic data.
| Kadri, Sameer S; Adjemian, Jennifer; Lai, Yi Ling et al. (2018) Difficult-to-Treat Resistance in Gram-negative Bacteremia at 173 US Hospitals: Retrospective Cohort Analysis of Prevalence, Predictors, and Outcome of Resistance to All First-line Agents. Clin Infect Dis : |
