In a recent study published in JAMA Network Open, researchers estimated the association between the receipt of appropriate initial empirical antibiotic treatment and in-hospital death among individuals hospitalized with BSI (bloodstream infections).
Bloodstream infections have been identified as an important health concern across the globe and result in patient deaths within a short period. Initial and sufficient empirical antibiotic treatment is the key to patient survival for individuals with bloodstream infections and sepsis. However, data on BSI epidemiology and the usage of suitable empirical antimicrobial treatment are limited.
About the study
In the present retrospective study, researchers examined the pathogenic organism characteristics, antibiotic resistance profiles, and within-hospital deaths related to receiving suitable empirical antibiotic treatment among BSI patients in the United States (US).
The study was conducted between 2016 and 2020 by utilizing the Premier Inc (Premier health database) data of 32,100 adults with BSIs across 183 hospitals in the US with positive initial blood cultures in the hospitalization period and those who were provided empirical antibiotic treatment (≥1.0 novel systemic antibiotic in the initial two days of collecting blood samples).
The team excluded individuals with polymicrobial-type infections. Data concerning sex, age, ethnicity, race, comorbid conditions, payer category, hospital transfer, nosocomial infections, the primary site of infection, and hospital characteristics such as bed size, hospital teaching status, and hospital location were obtained. In addition, data on disease severity markers such as intensive care unit (ICU) admissions, vasopressor usage, dialysis, mechanical ventilation requirements, total platelet count, total bilirubin, and creatinine were obtained.
Primary sites of infection were identified based on the ICD-10 (international statistical classification of diseases and related health problems, tenth revision) diagnosis codes. Multilevel multivariate logistic regression modeling was performed to assess the relationship between the receival of suitable early empirical antibiotic treatment and within-hospital deaths among individuals with GNR (gram-negative-type rods), GPC (gram-positive-type cocci), or Candida infections.
In addition, the team evaluated the incidence of infections with resistant pathogens such as MRSA (methicillin-resistant Staphylococcus aureus), VRE (vancomycin-resistant Enterococcus), ESBL (extended-spectrum beta-lactamase) gram-negative pathogens, CRE (carbapenem-resistant Enterobacterales) who were resistant to meropenem, imipenem, ertapenem, or doripenem), and CTX-R0 (ceftriaxone-resistant gram-negative organisms) such as ESBL and Pseudomonas aeruginosa.
The team considered BSI caused by coagulase-negative Staphylococcus organisms as contaminants. Further, sensitivity analyses were performed, wherein the team examined appropriate empirical therapy usage for every pathogen and every resistant pathogenic organism, eliminating disease severity covariates from the analyses.
Out of 32,100 individuals with bloodstream infections who received novel empirical antibiotics with an average age of 64 years, 55% and 70% were men, and Whites of non-Hispanic ethnicities, respectively. An in-hospital death rate of 14% was observed. Escherichia coli (58%) and S. aureus (32%) were the most commonly observed pathogens, followed by Klebsiella (21%), Proteus (nine percent), and Pseudomonas aeruginosa (eight percent).
Out of 32,100 individuals, CTX-RO, ESBL, and CRE were isolated in 10%, three percent, and one percent, respectively. The frequency of MRSA in all BSIs was 14%, and the frequency of MRSA among individuals with Staphylococcus aureus infections was 44%. The prevalence of VRE among all bloodstream infections was one percent, and 47%, 53%, and 1.0% of patients had positive reports for GNRs, GPC, and Candida, respectively.
The common sites of infection were genitourinary and pulmonary across the groups. Congestive cardiac failure, diabetes, and renal diseases were the most commonly observed comorbid conditions. Vancomycin, cephalosporin, and penicillin were the most commonly used initial empirical treatment across the groups. Over 94% of BSIs were not hospital-acquired infections.
Crude percentages of suitable empirical antibiotic treatments were 94%, 97%, and 65% for GNR, GPC, and Candida infections, respectively. The percentages of suitable antibiotic treatment usage for resistant pathogens were 60% and 55% for VRE and CRE, respectively. Acinetobacter had the least percent of suitable empirical antimicrobial therapy (65%) among non-resistant GNRs.
For Candida species, in 102 individuals who received inappropriate empirical antimicrobial therapy, the proportion who did not receive empirical antifungal therapy was 92%. Compared with unsuitable empirical antibiotic treatment, the receipt of suitable empirical antibiotic treatment was related to lesser in-hospital mortality risks for GNR, GPC, and Candida groups, with aOR (adjusted odds ratio) values of 0.5, 0.6, and 0.5, respectively.
Similar results were obtained in the sensitivity analyses after excluding disease severity covariates, with aOR values of 0.5, 0.7, and 0.4 for the GNR, GPC, and Candida species, respectively. In the sensitivity analysis by pathogen type, the in-hospital mortality risks associated with suitable empirical therapy were lower for individuals infected with Escherichia coli, Klebsiella, CTX-RO, S. aureus, MRSA, and Enterococcus, with aOR values of 0.6, 0.5, 0.6, 0.4, 0.5, and 0.5, respectively.
Overall, the study findings showed that among patients hospitalized with BSIs, the receipt of suitable empirical antimicrobial treatment was related to lesser in-hospital deaths. It is essential for clinical professionals to carefully select empirical antimicrobials for improving outcomes among BSI patients.