Integrons, a predictive biomarker for antibiotic resistance in acute sepsis: the IRIS study

Background: Considering the increase in MDR Gram-negative bacteria (GNB), the choice of empirical antibiotic therapy is challenging. In parallel, use of broad-spectrum antibiotics should be avoided to decrease antibiotic selection pressure. Accordingly, clinicians need rapid diagnostic tools to narrow antibiotic therapy. Class 1-3 integrons, identified by intI1-3 genes, are genetic elements that play a major role in antibiotic resistance in GNB. Objectives: The objective of the IRIS study was to evaluate the negative and positive predictive values (NPVs and PPVs, respectively) of intI1-3 as markers of antibiotic resistance. Methods: The IRIS study was an observational cross-sectional multicentre study that enrolled adult subjects with suspected urinary tract or intra-abdominal infections. intI1-3 were detected directly from routinely collected biological samples (blood, urine or intra-abdominal fluid) using real-time PCR. A patient was considered `MDR positive' if at least one GNB, expressing acquired resistance to at least two antibiotic families among b-lactams, aminoglycosides, fluoroquinolones and/or co-trimoxazole, was isolated from at least one biological sample. Results: Over a 2 year period, 513 subjects were enrolled and 409 had GNB documentation, mostly Enterobacterales. intI1 and/or intI2 were detected in 31.8% of patients and 24.4% of patients were considered `MDR positive'. The NPV of intI1 and/or intI2 as a marker of acquired antibiotic resistances was estimated at 92.8% (89.1%-95.5%). The NPVs for first-line antibiotics were all above 92%, notably >96% for resistance to third-generation cephalosporins. Conclusions: The IRIS study strongly suggests that the absence of intI1 and intI2 in biological samples from patients with GNB-related infections is predictive of the absence of acquired resistances.