Search

A multicenter, retrospective study of patients infected with carbapenem-resistant Pseudomonas aeruginosa who were treated with ceftolozane/tazobactam was performed. Among 35 patients, pneumonia was the most common indication and treatment was successful in 26 (74%). Treatment failure was observed in all cases where isolates demonstrated ceftolozane-tazobactam minimum inhibitory concentrations ≥8 μg/mL.

Background Ceftolozane/tazobactam (C/T) is a novel cephalosporin/beta-lactamase inhibitor combination developed for use against multidrug-resistant (MDR) Gram-negative infections, particularly Pseudomonas aeruginosa (PA). C/T is approved for complicated urinary tract and intraabdominal infections as well as hospital-acquired/ventilator-associated bacterial pneumonias. However, comprehensive clinical characterization of patients treated with C/T in non-FDA-approved indications is limited. Methods Patients ≥18 years who received C/T for ≥48 hours while hospitalized in 9 acute care centers in Houston, TX from January 2016 through September 2018 were included. Demographic, microbiologic, treatment and clinical outcome data were retrospectively collected by chart review. In patients who received multiple inpatient courses of C/T, only the first course with C/T was assessed. Results 210 patients met inclusion criteria: 58% were non-white, 35% were female and 13% were immunocompromised. Median age was 61 years (IQR, 48 to 69). Median Charlson comorbidity index was 5 (IQR, 2 to 6). At the onset of the index episode, a significant proportion of patients required intensive care unit admission (44%), mechanical ventilation (37%) and pressor support (22%). Respiratory sources were the most common (50%) followed by urine (15%). Positive cultures were documented in 93% of the cases and PA was found in 86%. Majority (95%) of PA which were MDR. C/T use was guided by susceptibility testing of the index isolate in ca. 52%. In 5.7% of cases, C/T was used to escalate therapy without any documented C/T-susceptible organism. Half (51%) of the cohort received initial dosing appropriate for renal function while 36% receiving a lower than recommended dose. Clinical success (i.e., recovery from infection-related signs and symptoms) occured in 77%. The in-hospital mortality rate in our cohort was 15% with 26 of 31 deaths deemed infection-related. Conclusion We report a large multicenter observational cohort that received C/T. A 77% clinical success with the use of C/T was documented. These data support the use of C/T in critically ill patients infected with MDR PA. Disclosures William R. Miller, MD, Entasis Therapeutics (Scientific Research Study Investigator)Merck (Grant/Research Support)Shionogi (Advisor or Review Panel member) Laura A. Puzniak, PhD, Merck (Employee) Cesar A. Arias, MD, MSc, PhD, FIDSA, Entasis Therapeutics (Scientific Research Study Investigator)MeMed (Scientific Research Study Investigator)Merck (Grant/Research Support)

Background Coagulase-negative staphylococci, including Staphylococcus epidermidis, are the most common cause of bloodstream infection in cancer patients. Linezolid resistance is increasingly identified in S. epidermidis, but whether such resistance alters the clinical course of S. epidermidis infections is unknown. The purpose of this study was to assess the clinical impact of linezolid resistance in leukemia patients with S. epidermidis bloodstream infection. Methods This was a retrospective, single-center cohort study of all adult leukemia patients with S. epidermidis bacteremia treated with empiric linezolid between 2012 and 2015. The primary end point was adverse clinical outcome on day 3, defined as a composite of persistent bacteremia, fever, intensive care unit admission, or death. Fourteen- and 30-day mortality were also assessed. Results Eighty-two unique leukemia patients with S. epidermidis were identified. Linezolid resistance was identified in 33/82 (40%). Patients with linezolid-resistant S. epidermidis were significantly more likely to have persistent bacteremia (41% vs 7%; adjusted relative risk [aRR], 5.15; 95% confidence interval [CI], 1.63–16.30; P = .005); however, adverse short-term clinical outcomes overall were not more common among patients with linezolid-resistant S. epidermidis (61% vs 33%; aRR, 1.46; 95% CI, 0.92–2.32; P = .108). No differences were observed in 14- or 30-day mortality. Conclusions Leukemia patients with linezolid-resistant S. epidermidis bacteremia who were treated with linezolid were significantly more likely to have persistent bacteremia compared with those with linezolid-sensitive isolates. Interventions to limit the clinical impact of linezolid-resistant S. epidermidis are warranted.

Background Staphylococcus epidermidis is a leading cause of healthcare-associated bacteremia, but the clinical course of S. epidermidis isolated from blood cultures ranges from contamination to serious deep seated infections. We tested the hypothesis that the genetic characteristics of S. epidermidis isolated from blood cultures are significantly associated with disease severity using whole genome sequencing (WGS). Methods We performed WGS of 163 S. epidermidis isolates from a large prospective multicenter, randomized control trial that assessed the safety and efficacy of algorithm-based treatment of patients with staphylococcal bacteremia (ClinicalTrials.gov #NCT01191840). Patient’s infection types were divided into simple (including possible contamination), uncomplicated and complicated bacteremia. Failure was defined as persistent signs and symptoms of infection at 72 hours, persistent bacteremia at 7 days, relapse, complications or overall mortality within 28 days of therapy. WGS was performed using Illumina Miseq, followed by in silico multi-locus sequence type identification and phylogenomic analysis using kSNP. Results Of the 163 isolates analyzed, 39 sequence types (STs) were identified with ST2 and ST5 isolates being most frequently identified (21% and 20%, respectively). ST2 and ST5 isolates were significantly more likely associated with uncomplicated and complicated infections rather than simple bacteremia (P = 0.01 by χ2 test). By multivariate regression analysis, having an ST2 or ST5 S. epidermidis bacteremia was an independent predictor of a complicated infection (odds ratio 4.28, 95% CI 1.36–13.42). Using WGS based branching points rather than sequence typing allowed for additional strengthening of the association of S. epidermidis genetic clustering and clinical infection types (figure). Although there was no significant difference in failure rates among patients infected with different STs, ST2/ST5 strains were significantly more likely to cause relapsing infections (85.7%, P = 0.02). Conclusion WGS could offer a prognostic tool in the management of S. epidermidis bacteremia. ST2 and ST5 strains may help predict a complicated bacteremia course which would warrant appropriate antimicrobial therapy. Disclosures T. L. Holland, Basilea: Consultant, Consulting fee. Motif Bio: Consultant and Scientific Advisor, Consulting fee. Theravance: Consultant, Speaker honorarium. Genentech: Consultant, Consulting fee. V. G. Fowler Jr., Debiopharm: Consultant, Consulting fee. Durata: Consultant, Consulting fee. Merck: Consultant and Scientific Advisor, Consulting fee. Cerexa/Actavis/Allergan: Grant Investigator, Grant recipient. Pfizer: Consultant and Grant Investigator, Consulting fee and Grant recipient. Advanced Liquid Logics: Grant Investigator, Grant recipient. NIH: Grant Investigator, Grant recipient. MedImmune: Consultant and Grant Investigator, Consulting fee and Grant recipient. Basilea: Consultant and Grant Investigator, Consulting fee and Grant recipient. Karius: Grant Investigator, Grant recipient. Contrafect: Consultant and Grant Investigator, Consulting fee and Grant recipient. Regeneron: Grant Investigator, Grant recipient. Genentech: Consultant and Grant Investigator, Consulting fee and Grant recipient. Affinergy: Consultant and Grant Investigator, Consulting fee and Grant recipient. Locus: Grant Investigator, Grant recipient. Medical Surface, Inc.: Grant Investigator, Grant recipient. Theravance: Consultant, Consulting fee and Speaker honorarium. Green Cross: Consultant, Speaker honorarium. Grifols: Consultant, Consulting fee. xBiotech: Consultant, Consulting fee. Achaogen: Consultant, Consulting fee. Medicines Co: Consultant, Consulting fee. Novartis: Consultant, Consulting fee. Novadigm: Consultant, Consulting fee. Bayer: Consultant, Consulting fee. Cubist: Consultant, Consulting fee. Debiopharm: Consultant, Consulting fee. Durata: Consultant, Consulting fee. I. Raad, The University of Texas MD Anderson Cancer Center: Shareholder, Licensing agreement or royalty. The Unversity of Texas MD Anderson Cancer Center: Shareholder, Dr. Raad is a co-inventor of the Nitroglycerin-Citrate-Ethanol catheter lock solution technology which is owned by the University of Texas MD Anderson Cancer Center (UTMDACC) and has been licensed to Novel Anti-Infective Technologies LLC, in which UTMDACC and Licensing agreement or royalty.

Background Outcomes are improved with ceftazidime–avibactam (CZA) compared with polymyxin-based regimens (PBR) for carbapenemase-producing carbapenem-resistant Enterobacteriaceae. It is unclear whether this finding is true in non-carbapenemase (non-CP) producing CRE. The purpose of this study was to compare the efficacy and safety of CZA-based and PBR for CRE bacteremia in cancer patients with a high prevalence of non-CP CRE. Methods Adult cancer patients with first occurrence of CRE (i.e., meropenem non-susceptible) bacteremia treated with either CZA or PBR as directed therapy were included. Day 14 integrated benefit-risk outcomes based on desirability of outcome ranking (DOOR; (1) cured and discharged home, (2) cured and hospitalized, (3) cured and hospitalized with renal failure, (4) not cured, (5) dead) were used. DOOR is a recently developed statistical approach designed to unify important patient and clinician outcomes. Inverse probability of treatment weighted (IPTW) ordered logistic regression was used to model the odds of moving down ranked DOOR categories (i.e., having a worse outcome). The probability of a patient treated with CZA or a PBR having a worse DOOR category was also calculated. IPTW logistic regression was used to model the odds of 14-day mortality. Results 43 patients (CZA, n = 24; PBR, n = 19) with similar demographics and relative illness were included. Klebsiella pneumoniae (n = 21) and Escherichia coli (n = 16) were most common. 16/43 (37%) were CP CRE, 19/43 (44%) were non-CP CRE, and the remainder were unknown. The probability of a better DOOR for patients treated with CZA was 58% (95% CI 53% - 62%). Patients treated with CZA had an 81% reduction in IPTW-adjusted odds of a worse DOOR (OR 0.19, 95% CI 0.05 – 0.76; P = 0.02). 14-day mortality was 2/24 (8%) for patients receiving CZA vs. 5/19 (26%) for patients treated with PBR (IPTW-adjusted OR 0.12, 95% CI 0.02 – 0.82, P = 0.03). Conclusion These data suggest that CZA-based treatment, compared with PBR, has a superior integrated benefit-risk profile for the treatment of CRE bacteremia in cancer patients with a high burden of non-CP CRE. These findings build upon available data and suggest that CZA is preferred to PBR for CRE with heterogenous resistance mechanisms. Disclosures Samuel L. Aitken, PharmD, Melinta Therapeutoics: Grant/Research Support, Research Grant; Merck, Sharpe, and Dohme: Advisory Board; Shionogi: Advisory Board.

Background S. maltophilia infections are increasingly common in patients with acute myeloid leukemia (AML) and associated with high mortality. S. maltophilia is a frequent colonizer in AML patients but relatively little is known about factors that drive S. maltophilia infection. We sought to evaluate the utility of cumulative antibiotic use and the microbiome as predictors of S. maltophilia infection in AML patients receiving remission induction chemotherapy (RIC). Methods We performed a subanalysis of a prospective, observational cohort study between September 2013 and August 2015 of adult patients with AML receiving RIC. In this study, fecal and oral microbiome samples collected every 96 hours from the start of RIC until neutrophil recovery were assessed for the relative abundance of S. maltophilia via 16s rRNA quantitation. The primary outcome, microbiologically proven S. maltophilia infection, was analyzed using a time-varying Cox proportional hazards model accounting for S. maltophilia relative abundance and cumulative antibiotic exposure. Patients were censored at neutrophil recovery or death. Results 90 patient were included; of whom, 8 (9%) developed S. maltophilia infection (pneumonia, n = 6; skin/soft-tissue infection, n = 2). 4/8 (50%) patients were bacteremic. 7/8 (88%) patients with S. maltophilia infection had detectable oral 16s oral reads mapping to S. maltophilia vs 22/82 (27%) without infection (P < 0.01). An oral S. maltophilia relative abundance of 36% predicted infection (sensitivity: 96%, specificity 93%, likelihood ratio +: 17.08). No association of S. maltophilia infection with the fecal relative abundance was seen. Cumulative meropenem exposure was associated with increased infection risk (hazard ratio [HR] 1.17, 95% CI 1.01 – 1.35, P = 0.03), while levofloxacin was associated with decreased infection risk (HR 0.83, 95% CI 0.66 – 1.04, P = 0.10). Conclusion The oral microbiome may play an important role in S. maltophilia pathogenesis in AML patients. Cumulative antibiotic exposure likely modifies S. maltophilia infection risk. These data suggest that real-time molecular monitoring of the oral cavity for S. maltophilia in AML patients could identify patients at high risk for S. maltophilia infection and improve targeted therapy. Disclosures Samuel L. Aitken, PharmD, Melinta Therapeutoics: Grant/Research Support, Research Grant; Merck, Sharpe, and Dohme: Advisory Board; Shionogi: Advisory Board.

Background Since 1970 group B Streptococcus (GBS) has been a frequent cause of sepsis or meningitis in young infants. Capsular polysaccharide type V was first recognized in 1990 and has increased to the point where it now causes ~15% of GBS infections. GBS type V strains are almost entirely sequence type 1 (ST1) in adult infections. To understand the emergence of type V GBS, we compared infant strains before 1990 to more contemporary isolates from young infants and adults. Methods Thirty-five strains isolated from blood or CSF of infants Results The majority (29/35) of Houston young infant strains were ST1. Type V GBS strains isolated prior to 1990 were more likely to be of ST-2 or ST-26 (5/10) compared with those from 1990 or later (24/25 and 14/14 CDC infant invasive type V). Tetracycline resistance was identified in 83% (29/35) while macrolide resistance (MR) occurred in only 23% (8/35) of the strains. Compared with early neonatal isolates, MR was significantly more frequent among contemporary neonatal (12/14, 86%, P < 0.0001) and adult (502/710, 71%, P < 0.0001) ST1 GBS. Phylogenetic analysis showed two distinct clades defined, in part, by MR. A high-frequency MR (340/360, 94%) clade was defined by the presence of erm(B) on Tn3872 while the low-frequency MR clade (159/350, 45%) was more diverse in mobile elements contributing to MR. The majority (27/29) of early neonatal ST1 GBS strains were observed in the low-frequency MR clade. Conclusion Infant invasive disease due to type V GBS before 1990 consisted of more diverse STs but is now almost exclusively ST1. Differences in the frequency of MR between early neonatal and contemporary type V ST1 GBS suggest MR may, at least in part, have driven the expansion of type V ST1 GBS. Disclosures All Authors: No reported Disclosures.

Background Ceftazidime–avibactam (CZA) is a novel β-lactam / β-lactamase inhibitor with in vitro activity against multidrug-resistant Gram-negatives, including those harboring CMY-2 enzymes. Treatment-emergent resistance to CZA has been described in KPC-producing Klebsiella pneumoniae but has not been described in non-carbapenemase-producing, carbapenem-resistant Enterobacteriaceae (CRE). Methods A patient with an intra-abdominal infection due to a carbapenem-resistant E. coli (ertapenem MIC 16 µg/mL; meropenem MIC 2 µg/mL; CZA MIC 2 µg/mL; carbapenemase negative) was treated with CZA. On day 48 of therapy, a CZA resistant, carbapenem-sensitive E. coli was identified from abdominal drainage (CZA MIC ≥256 µg/mL; meropenem MIC 0.19 µg/mL). Illumina MiSeq whole-genome sequencing (WGS) was performed on both isolates to identify potential resistance mechanisms. The ResFinder database was used to identify known β-lactamase enzymes, and in silico modeling of β-lactamase structure was assessed. Results WGS revealed that both isolates were ST410 E. coli, with the sole difference in β-lactam resistance determinants between the two being a novel CMY β-lactamase harbored on an Inc1-type conjugative plasmid in the second isolate. The novel CMY has 4 amino acid substitutions relative to CMY-2: A134E, Q140K, V231S, and N366Y. The V231S substitution is found in CMY-42 and has previously been associated with increased ceftazidime hydrolysis. The remaining three substitutions have not previously been identified. Previous studies have identified that substitutions at position 366 influence the rate of ceftazidime hydrolysis rate. Preliminary protein structure analysis suggests that positions 140 and 366 are in the active site. No other differences in β-lactam resistance determinants were identified between the first and second isolates. Conclusion To our knowledge, we have identified the first case of CMY-associated CZA resistance. Given the widespread and transferrable nature of CMY enzymes, this finding raises concern for additional cases of resistance with increasing usage of CZA. Further analysis is needed to identify the mechanism by which this enzyme confers CZA resistance. Disclosures Samuel L. Aitken, PharmD, Melinta Therapeutics: Grant/Research Support, Research Grant; Merck, Sharpe, and Dohme: Advisory Board; Shionogi: Advisory Board.

Background Fluoroquinolones (FQs) are routinely used as antimicrobial prophylaxis in leukemia patients receiving chemotherapy to prevent Pseudomonas aeruginosa infections. Patients who are intolerant to FQs may receive cefpodoxime (CPD) or other agents. How FQ use affects the resistance profile and epidemiology of breakthrough P. aeruginosa infections is unknown. To determine this, we performed a whole-genome sequencing (WGS)-driven epidemiologic study of leukemia patients with P. aeruginosa bloodstream infections. Methods All adult (age > 17 years) inpatients with leukemia and a first episode of monomicrobial P. aeruginosa bloodstream infection were included. Clinical data were extracted from the electronic medical record. Isolates were sequenced using an Illumina NextSeq and phylogenomics was performed using an in-house analysis pipeline consisting of Bowtie2, SAMtools and bcftools. Results 110 patients were included and most had a diagnosis of acute myeloid leukemia (n = 66). Twenty (18%) patients received FQ prophylaxis, 56 (54%) received CPD, and the remaining 34 (31%) received other agents. 9 (8%) isolates were multidrug-resistant (MDR). MDR was more common in those receiving FQ prophylaxis (20% vs 6%, P = 0.06). 76 sequence types (STs) were represented with ST235 (n = 8) being most common followed by ST244 (n = 7). ST235 strains were genetically distinct, but closely related (>10 but < 250 SNPs) in comparison to other STs. 2 ST244 strains were genetically identical despite being isolated 4 months apart, suggesting horizontal transmission. MDR was more common among ST235 isolates compared with other STs (38% vs 6%, P = 0.02). ST235 strains were more common in patients receiving FQ vs other prophylaxis (20% vs 4%, P = 0.04). 1 ST244 isolate harbored a VIM-2 β-lactamase. In 20 FQ-resistant isolates, 80% had mutations in either parC (S87L) or gyrA (T83I) and 50% had both. FQ-resistance mutations were more common in FQ recipients (50% vs 8%, P < 0.01). Conclusion Most P. aeruginosa infections occurred in non-FQ recipients, while MDR P. aeruginosa infections were more common in FQ recipients. These data suggest that decisions on empiric treatment of patients with P. aeruginosa bacteremia must take antimicrobial prophylaxis history into account. Disclosures Samuel L. Aitken, PharmD, Melinta Therapeutoics: Grant/Research Support, Research Grant; Merck, Sharpe, and Dohme: Advisory Board; Shionogi: Advisory Board.

Hfq is a post-transcriptional regulator that plays a key role in bacterial gene expression by binding AU-rich sequences and A-tracts to facilitate the annealing of sRNAs to target mRNAs and to affect RNA stability. To understand how Hfq from the Gram-positive bacterium Staphylococcus aureus (Sa) binds A-tract RNA, we determined the crystal structure of an Sa Hfq-adenine oligoribonucleotide complex. The structure reveals a bipartite RNA-binding motif on the distal face that is composed of a purine nucleotide-specificity site (R-site) and a non-discriminating linker site (L-site). The (R-L)-binding motif, which is also utilized by Bacillus subtilis Hfq to bind (AG)(3)A, differs from the (A-R-N) tripartite poly(A) RNA-binding motif of Escherichia coli Hfq whereby the Sa Hfq R-site strongly prefers adenosine, is more aromatic and permits deeper insertion of the adenine ring. R-site adenine-stacking residue Phe30, which is conserved among Gram-positive bacterial Hfqs, and an altered conformation about β3 and β4 eliminate the adenosine-specificity site (A-site) and create the L-site. Binding studies show that Sa Hfq binds (AU)(3)A ≈ (AG)(3)A ≥ (AC)(3)A > (AA)(3)A and L-site residue Lys33 plays a significant role. The (R-L) motif is likely utilized by Hfqs from most Gram-positive bacteria to bind alternating (A-N)(n) RNA.

Background Carbapenems are the treatment of choice for bacteremia caused by extended spectrum β-lactamase producing Enterobacteriaceae (ESBL-E). The emergence carbapenem resistance (CR) in ESBL-E isolates has been described; however, the rate of such resistance in clinical settings is unknown. We describe the frequency and mechanisms of CR in recurrent ESBL-E bacteremia at an NCI-designated cancer center. Methods We performed a prospective whole genome sequencing (WGS) study and retrospective cohort review of adult (age ≥18 years) patients with ESBL-E bacteremia between January 2015 and July 2016. Recurrent bacteremia was defined as identification of the same organism in blood culture at any time following initial successful treatment. CR was defined as resistance to meropenem. Carbapenemase production was assessed in the microbiology laboratory using Carba-NP. Available paired isolates underwent WGS via Illumina HiSeq for assessment of clonality and identification of CR mechanisms. Results One hundred and sixteen patients with ESBL-E bacteremia were identified. E. coli was the most common organism (86%), followed by K. pneumoniae (12%), and K. oxytoca (2%). Recurrent bacteremia was identified in 17 (15%) patients (E. coli [n = 15], K. pneumoniae [n = 2]). Of these, 6 (35%) were CR and 5/6 (83%) were Carba-NP negative. All six recurrent CR isolates occurred in patients with leukemia. Five isolate pairs were available for WGS. In four of five pairs (three E. coli, one K. pneumoniae), CR emerged from the same strain causing the original infection; one recurrence was caused by a distinct E. coli with a OXA-48-like gene. Compared with parental strains, CR E. coli contained deletions in porin-encoding genes and had increased mapping depth for genes encoding CTX-M ESBLs. The K. pneumoniae was Carba-NP negative with no identifiable CR mechanism. Conclusion Emergence of CR following treatment for ESBL-E bacteremia was seen only in leukemia patients and was primarily due to porin loss and amplification of ESBL genes, rather than acquisition of exogenous carbapenemases. These are the first clinical data describing the molecular mechanism of ESBL-E transformation to CR. These data serve as the basis for future studies of antimicrobial stewardship interventions to limit the emergence of CR in ESBL-E. Disclosures S. L. Aitken, Shionogi: Scientific Advisor, Consulting fee. Medicines Company: Scientific Advisor, Consulting fee. Merck & Co.: Scientific Advisor, Consulting fee.

Background Patients undergoing chemotherapy for leukemia are at high risk for infection and routinely receive antibiotic prophylaxis. The types of breakthrough bloodstream infection (BSI) based on choice of prophylaxis is not well-characterized. Here, we describe antibiotic prophylaxis patterns and the influence of antibiotic choice on BSI epidemiology in leukemia patients presenting to the emergency center (EC) with neutropenic fever (NF). Methods This was a retrospective chart review of patients with leukemia and NF (absolute neutrophil count [ANC] 7 days in duration and the median ANC at presentation was 0 cells/mm3 (range: 0–490 cells/mm3). Most patients received LEV (42%) followed by CIP (27%), CEF (25%), and ACL (6%). Forty-seven of 284 patients presented with Gram-negative BSI (16%) and 36 (13%) had Gram-positive BSI. Rates of common organisms causing BSI are presented in Table 1. Conclusion In leukemia patients with NF presenting to the EC, rates of BSI differed significantly based on antibiotic prophylaxis choice, with P. aeruginosa BSI more common in patients receiving ACL and E. coli in patients receiving LEV. The epidemiology of breakthrough infections on different prophylactic agents may help guide empiric antibiotic choice. Table 1. Causative organisms of BSI. BSI Type LEV (n = 118) CIP (n = 77) CEF (n = 72) ACL (n = 17) P-value Any Gram-negative (n = 47) 21 (18) 7 (9) 13 (18) 6 (35) 0.05 E. coli (n = 25) 18 (15) 3 (4) 3 (4) 1 (6) 0.02 P. aeruginosa (n = 13) 2 (2) 1 (1) 6 (8) 4 (24)

Background We have previously shown that vancomycin-resistant E. faecium (VRE) with DAP MICs close to the breakpoint (4 µg/mL) harbor genetic changes associated with DAP resistance (DAP-R). Further, DAP MIC was a predictor of poor microbiological eradication in patients with VRE bacteremia treated with DAP. Furthermore, DAP-susceptible VRE isolates with DAP MIC of 3–4 µg/mL (Etest) were more likely to fail DAP therapy, independently of the DAP dose used. Here, we used WGS to determine whether mutations associated with DAP-R could predict DAP MICs. Methods We performed WGS to identify potential determinants of DAP-R in 80 E. faecium isolates (62 DAP-S and 18 DAP-R recovered from bloodstream and other infection sites) in diverse US geographical locations. Two modeling strategies were employed with the aim of increasing the robustness of our prediction strategy, (i) a logistic regression model approach to predict the probability of an isolate of exhibiting a DAP MIC of ≥ 3µg/dl based on the presence of relevant mutations, and (ii) a linear regression model to predict a single doubling dilution increase on DAP MIC in the presence or absence of mutations associated with DAP-R, after transforming MICs to a log2 scale. Statistical significance (P value) was set at