Search

Background: Identification of hospitalized patients with enteric multidrug-resistant organism (MDRO) carriage, combined with implementation of targeted infection control interventions, may help reduce MDRO transmission. However, the optimal surveillance approach has not been defined. We sought to determine whether daily serial rectal surveillance for MDROs detects more incident cases (acquisition) of MDRO colonization in medical intensive care unit (MICU) patients than admission and discharge surveillance alone. Methods: Prospective longitudinal observational single-center study from January 11, 2017, to January 11, 2018. Inclusion criteria were ≥3 consecutive MICU days and ≥2 rectal or stool swabs per MICU admission. Daily rectal or stool swabs were collected from patients and cultured for MDROs, including vancomycin-resistant Enterococcus (VRE), carbapenem-resistant Enterobacterales (CRE), third-generation cephalosporin-resistant Enterobacterales (3GCR), and extended-spectrum β-lactamase–producing Enterobacterales (ESBL-E) (as a subset of 3GCR). MDRO detection at any time during the MICU stay was used to calculate prevalent colonization. Incident colonization (acquisition) was defined as new detection of an MDRO after at least 1 prior negative swab. We then determined the proportion of prevalent and incident cases detected by daily testing that were also detected when only first swabs (admission) and last swabs (discharge) were tested. Data were analyzed using SAS version 9.4 software. Results: In total, 939 MICU stays of 842 patients were analyzed. Patient characteristics were median age 64 years (interquartile range [IQR], 51–74), median MICU length of stay 5 days (IQR, 3–8), median number of samples per admission 3 (IQR, 2–5), and median Charlson index 4 (IQR, 2–7). Prevalent colonization with any MDRO was detected by daily swabbing in 401 stays (42.7%). Compared to daily serial swabbing, an admission- and discharge-only approach detected ≥86% of MDRO cases (ie, overall prevalent MDRO colonization). Detection of incident MDRO colonization by an admission- or discharge-only approach would have detected fewer cases than daily swabbing (Figure 1); ≥34% of total MDRO acquisitions would have been missed. Conclusions: Testing patients upon admission and discharge to an MICU may fail to detect MDRO acquisition in more than one-third of patients, thereby reducing the effectiveness of MDRO control programs that are targeted against known MDRO carriers. The poor performance of a single discharge swab may be due to intermittent or low-level MDRO shedding, inadequate sampling, or transient MDRO colonization. Additional research is needed to determine the optimal surveillance approach of enteric MDRO carriage.Funding: NoDisclosures: None

Background MDRO detection in colonized patients may be intermittent for reasons that are incompletely understood. We examined temporal patterns of gut MDRO colonization after initial MDRO detection by rectal swab screening, and determined the relationship of culture positivity to the relative abundance of corresponding MDRO operational taxonomic units (OTUs) identified by 16S rRNA gene sequence analysis. Methods Rectal or fecal swabs were collected daily from MICU patients 1/11/2017-1/11/2018. First MICU admissions with ≥2 swabs and MICU stays ≥3 days were studied. Samples were cultured for vancomycin-resistant enterococci (VRE), carbapenem-resistant Enterobacteriaceae (CRE) and P. aeruginosa (CRPA), and extended-spectrum β-lactamase-producing (ESBL) Enterobacteriaceae by selective media. Resistance mechanisms were confirmed by phenotypic methods and/or PCR. Limit of detection was similar for different MDROs (24-52 CFU/sample). OTU categories corresponding to MDRO species were identified by taxonomy and BLAST. Multilevel regression models estimated the association between MDRO detection and relative abundance of the corresponding OTU. Results 796 unique patients with 3519 swabs were studied. Median (IQR) age was 64 (51-74) years, MICU length of stay was 5 (3-8) days, and number of samples-per-patient was 3 (2-5). Following initial MDRO detection, the probability of subsequent detection varied by MDRO type, and was highest for VRE and lowest for CRPA [Figure 1]. Within each sample, we found a significant association between MDRO detection and relative abundance of the corresponding OTU [Table 1]. In contrast, relative OTU abundance in the first sample with MDRO detection was not predictive of odds of future MDRO detection (p >0.05 for all comparisons). Carriage of >1 MDRO did not affect the odds of MDRO detection in later samples. Figure 1. Probability of Subsequent MDRO Detection after First Positive Varies by MDRO Type Table 1. Higher Mean Corresponding OTU Relative Abundance Within Each Sample is Associated with MDRO Detection Conclusion MDRO culture positivity in rectally colonized patients was correlated with relative abundance of the corresponding OTU in the same sample. Serial detection of different MDRO types was variable, possibly due to distinct microbial community dynamics of different MDRO types. Intermittent failure to detect MDROs could result in misattribution of MDRO acquisition, resulting in inappropriate investigation or intervention. Disclosures All Authors: No reported disclosures

Background An association between increased relative abundance of specific bacterial taxa in the intestinal microbiota and bacteremia has been reported in some high-risk patient populations. Methods We collected weekly rectal swab samples from patients at 1 long-term acute care hospital (LTACH) in Chicago from May 2015 to May 2016. Samples positive for Klebsiella pneumoniae carbapenemase-producing Klebsiella pneumoniae (KPC-Kp) by polymerase chain reaction and culture underwent 16S rRNA gene sequence analysis; relative abundance of the operational taxonomic unit containing KPC-Kp was determined. Receiver operator characteristic (ROC) curves were constructed using results from the sample with highest relative abundance of KPC-Kp from each patient admission, excluding samples collected after KPC-Kp bacteremia. Cox regression analysis was performed to evaluate risk factors associated with time to achieve KPC-Kp relative abundance thresholds calculated by ROC curve analysis. Results We collected 2319 samples from 562 admissions (506 patients); KPC-Kp colonization was detected in 255 (45.4%) admissions and KPC-Kp bacteremia in 11 (4.3%). A relative abundance cutoff of 22% predicted KPC-Kp bacteremia with sensitivity 73%, specificity 72%, and relative risk 4.2 (P = .01). In a multivariable Cox regression model adjusted for age, Charlson comorbidity index, and medical devices, carbapenem receipt was associated with achieving the 22% relative abundance threshold (P = .044). Conclusion Carbapenem receipt was associated with increased hazard for high relative abundance of KPC-Kp in the gut microbiota. Increased relative abundance of KPC-Kp was associated with KPC-Kp bacteremia. Whether bacteremia arose directly from bacterial translocation or indirectly from skin contamination followed by bloodstream invasion remains to be determined.

Background Among hospitalized patients, underlying variation in gut microbiota may confer differential risk for gut MDRO acquisition. Methods Rectal swab samples were collected from patients ≤2 days of MICU admission and then daily in the 27-bed MICU of an acute care hospital in Chicago, IL over 1 year. Patients were screened for MDRO colonization by selective culture (see Figure 1 for MDRO types); those with ≥2 swabs and MICU stays ≥3 days were studied. Bacterial 16S rRNA gene amplicon sequences were used for microbiota analysis. Medical records were reviewed. Results In preliminary analysis, 2,480 samples were collected from 627 patients who acquired 170 MDROs (Figure 1). Debilitation, co-morbidities, and certain medical devices were associated with MDRO acquisition, though admission MDRO status was not (table). While no interactions were detected between admission MDRO status and clinical predictors of MDRO acquisition, there were significant differences in gut microbiota composition at the time of MICU admission between patients colonized with an MDRO on admission and those not colonized (P < 0.001, using analysis of molecular variance (AMOVA) on distances). Therefore, we stratified our analysis by admission MDRO colonization status. For patients MDRO-colonized at admission, there were no significant differences in microbiota of patients who later did or did not acquire a new MDRO (AMOVA P-value = 0.32). For patients not MDRO-colonized on admission, there was a significant difference in microbiota of patients who later acquired an MDRO and those who did not (AMOVA P-value: 0.026). Differentially abundant operational taxonomic units (OTUs, based on 3% sequence difference) included OTUs classified as Anaerococcus and as other Clostridiales (higher in patients who remained uncolonized) and as Enterococcus (higher in patients who acquired an MDRO) (Figure 2). Diversity was also higher in patients who remained uncolonized (Wilcoxon test P-value: 0.035) (Figure 3). Conclusion Among patients not already colonized with an MDRO on admission, we identified gut microbiota differences associated with MDRO acquisition that could help explain patient-level variation in MDRO colonization resistance. Disclosures All Authors: No reported Disclosures.

Background Clinical culture results are sometimes used to estimate the burden of multidrug-resistant organisms (MDROs) in hospitals. The association between positive clinical culture results and prevalence of MDROs in the gut is incompletely understood. Methods Rectal swab or stool samples were collected daily from adult medical intensive care unit (MICU) patients and cultured for target MDROs using selective media between January 2017 and January 2018 at Rush University Medical Center, a 676-bed tertiary-care center in Chicago. Resistance mechanisms were confirmed by phenotypic methods and/or polymerase chain reaction. Clinical culture results during MICU stay were extracted from the hospital information system. Target MDROs included vancomycin-resistant Enterococci (VRE), carbapenem-resistant Enterobacteriaceae (CRE), extended-spectrum β-lactamase-producing Enterobacteriaceae (ESBL), carbapenem-resistant Pseudomonas aeruginosa (CRPA) and carbapenem-resistant Acinetobacter baumannii (CRAB). Patients with either a study or clinical culture positive for a target MDRO were analyzed. Results We collected 5,086 study samples from 1,661 unique admissions (1,419 patients) and included here data from 413 unique admissions (397 patients) with completed microbiologic analysis. Median (IQR) patient age was 65 (51–75) years and length of MICU stay was 3 (3–4) days. A total of 156 (37.8%) patients had a target MDRO detected from a study sample at any point; 57 (36.5%) patients had >1 MDRO detected. Overall prevalence of these MDROs was found to be 22.5% VRE, 6.5% CRE, 19.8% ESBL, 4.4% CRPA, and 0.7% CRAB. New MDRO acquisition was observed in 58 (14.6%) patients (figure). Once a target MDRO was detected in a study sample, 82.2% of subsequent study samples were positive for that MDRO. Only 13 (5.8%) patients had a positive clinical culture for any target MDRO during their MICU stay (table). Conclusion Clinical cultures capture only the tip of the resistance iceberg and alone are insufficient to guide MDRO-targeted prevention strategies. Universal infection prevention measures are an alternative that may be preferred in settings where overall prevalence of MDROs is moderate or high and patients may be colonized with >1 MDRO. Disclosures All authors: No reported disclosures.