Antibiotic resistance is a major cause of morbidity and mortality. However, a better understanding of the relationship between bacterial genetic markers, phenotypic resistance, and clinical outcomes is needed. We performed whole-genome sequencing on five medically important pathogens (
Acinetobacter baumannii
,
Enterobacter cloacae
,
Escherichia coli
,
Klebsiella pneumoniae
, and
Pseudomonas aeruginosa
) to investigate how resistance genes impact patient outcomes. A total of 168 isolates from 162 patients with Gram-negative infections admitted to Beilinson Hospital at Rabin Medical Center in Israel were included for final analysis. Genomes were analyzed for resistance determinants and correlated with microbiologic and clinical data. Thirty-day mortality from time of culture was 26.5% (43/162). Twenty-nine patients had carbapenem-resistant isolates (29/168, 17.2%), while 63 patients had multidrug-resistant isolates (63/168, 37.5%). Albumin levels were inversely associated with mortality and length of stay, while arrival from a healthcare facility and cancer chemotherapy predicted having a multidrug-resistant isolate. Sequencing revealed possible patient-to-patient transmission events.
bla
CTX-M-15
was associated with multidrug-resistance in
E. coli
(OR = 3.888,
P
= 0.023) on multivariate analysis. Increased
bla
OXA-72
copy number was associated with carbapenem-resistance in
A. baumannii
(
P
= 0.003) and meropenem minimum inhibitory concentration (
P
= 0.005), yet carbapenem-resistant isolates retained sensitivity to cefiderocol and sulbactam–durlobactam. RJX84154 was associated with multidrug-resistance across all pathogens (
P
= 0.0018) and in
E. coli
(
P
= 0.0024). Low albumin levels were associated with mortality and length of stay in this sample population.
bla
CTX-M-15
was correlated with multidrug-resistance in
E. coli
, and
bla
OXA-72
depth predicted meropenem minimum inhibitory concentration in
A. baumannii
. RJX84154 may play a role in multidrug-resistance.
IMPORTANCE
While there have been several studies that attempt to find clinical predictors of outcomes in patients hospitalized with bacterial infections, less has been done to combine clinical data with genomic mechanisms of antibiotic resistance. This study focused on a hospitalized patient population in Israel with infections due to medically important bacterial pathogens as a way to build a framework that would unite clinical data with both bacterial antibiotic susceptibility and genomic data. Merging both clinical and genomic data allowed us to find both bacterial and clinical factors that impact certain clinical outcomes. As genome sequencing of bacteria becomes both rapid and commonplace, near real-time monitoring of resistance determinants could help to optimize clinical care and potentially improve outcomes in these patients.
