Article information

Abstract

Trans Abstract

INTRODUCTION

Extended-spectrum β-lactamases (ESBL), first described in 1983, are inhibited by clavulanic acid, sulbactam, and tazobactam. In 1991, cefotaxime-hydrolyzing β-lactamase-Munich (CTX-M-1) ESBLs were identified [1]. The increasing prevalence of antimicrobial resistance, particularly ESBL-producing Enterobacteriaceae (ESBL-E), is a threat to health globally. The incidence of ESBL-E identified in bacterial cultures increased by 53% in the United States from 2012 to 2017 [2]. Between 2015 and 2019, the global prevalence of ESBL non-carbapenem-resistant Enterobacteriaceae (CRE) phenotypes was > 50% [3]. Since 2019, the prevalence of ESBL-E in South Korea has been ≥ 50% [4]. Because of the spread of ST131 with CTX-M-15, the community prevalence of ESBL has increased [5]. ESBL-E-induced bacteremia is associated with an increased hospitalization duration, adding to the clinical and economic burden of bacteremia. In a 2020 multicenter retrospective study of patients with ESBL or CRE Klebsiella pneumoniae and Escherichia coli (EC) bacteremia in South Korea, the 30-day mortality rate of ESBL-E bacteremia was 10.3%, with a treatment failure rate of 12.5%, and a total cost for medical expenses of up to $11,151 [6]. According to a guidance from the Infectious Diseases Society of America (IDSA), carbapenems are the preferred treatment options for pyelonephritis and complicated urinary tract infection caused by ESBL-E [7]. In vitro, piperacillin-tazobactam (PIP/TAZ) is effective in the treatment of urinary tract infections caused by ESBL-EC [8]. The rate of ESBL-EC has increased in community-acquired urinary tract infections [9], and although carbapenems are recommended for ESBL-E bacteremia, it is important to note that the rate of CRE continues to increase worldwide in the coronavirus disease 2019-endemic era [10]. Therefore, we evaluated the clinical outcomes of non-carbapenem drugs in patients with ESBL-E bacteremia.

MATERIALS AND METHODS

Study design

We conducted a retrospective case-control study involving adult patients with documented ESBL-E bacteremia using the electronic medical records of a teaching hospital. Between January 1, 2021 and December 30, 2021, 527 adults (≥ 18 years of age) with Enterobacteriaceae bacteremia were enrolled in the study. Patients were excluded if they had been infected with the same species (or ESBL-E) 3 months prior to the current positive result or if two or more species were identified in a specimen, including blood. Based on the exclusion criteria, 152 patients with ESBL-E bacteremia were included in the study (Fig. 1). This study was approved by the Institutional Review Board of Jeju National University Hospital (No. 2021-03-013-001).

Figure 1.

Study population. Jan., January; Dec., December; n, number; ESBL, extended-spectrum β-lactamase.

The patients were divided into two groups according to whether they received non-carbapenem or carbapenem therapy. The rates of treatment failure, 30-day mortality and microbiologic failure, and the duration of hospitalization and of antimicrobial therapy were compared between the two groups. Antimicrobial susceptibility testing and phenotypic identification of ESBL-E were performed using the Vitek 2 system (BioMérieux, Marcy-l'Étoile, France). ESBL-E was defined as the presence of ESBL positivity or resistance to cefotaxime.

RESULTS

Clinical characteristics of patients with ESBL-E bacteremia

A total of 118 patients with ESBL-E bacteremia were included in the study; their clinical characteristics are listed in Table 1. Among the patients, 54 (45.8%) were males, with a mean age of 71.1 ± 11.6 years. The overall 30-day inpatient mortality rate was 16.1%. Of the patients, 54 and 64 were classified into the non-carbapenem group (NCG) and carbapenem group (CG), respectively. The proportion of male patients was higher in the CG than in the NCG (54.7% vs. 35.2%, respectively; p= 0.042). The mean age in the NCG showed a trend toward being higher than in the CG (79.4 ± 12.3 vs. 75.0 ± 12.5, respectively; p= 0.058). There were no significant differences in the PBS or other baseline clinical characteristics between the two groups. Urinary tract infection was the most prevalent primary infection type in both of the groups. The most frequently isolated strain was EC (74.1% in the NCG vs. 68.8% in the CG; p= 0.122). Hospitalization duration was significantly longer in the CG than the NCG.

Table 1.

Clinical characteristics of the patients with bacteremia caused by ESBL-producing Enterobacteriaceae bacteremia in the carbapenem and non-carbapenem groups

Primary and secondary outcomes

There was no significant difference in primary and secondary outcomes between the CG and NCG (Fig. 2). At 30 days, there was no significant difference in the primary outcome of treatment failure between the NCG and CG (16.7% vs. 18.8%; p= 0.65) (Fig. 2A). The secondary outcome of 30-day mortality occurred in eight and 11 patients in the NCG and the CG, respectively (14.8% vs. 17.2%; p= 0.63) (Fig. 2B). A Cox regression analysis showed no significant difference in the rate of treatment failure (HR, 1.32; 95% CI, 0.55-3.17; p= 0.54) (Fig. 3A) or 30-day mortality (HR, 1.75; 95% CI, 0.51-6.05; p= 0.38) (Fig. 3B) between the NCG and CG.

Figure 2.

Survival of patients with extended-spectrum β-lactamase (ESBL)-Enterobacteriaceae bacteremia in the carbapenem and non-carbapenem groups by Kaplan-Meier analysis. (A) Treatment failure at 30 days, (B) 30-day mortality rate.

Figure 3.

Cox regression analysis of patients with extended-spectrum β-lactamase (ESBL)-Enterobacteriaceae bacteremia in the the carbapenem and non-carbapenem groups. (A) Treatment failure at 30 days, (B) 30-day mortality rate. HR, hazard ratio; CI, confidence interval.

Risk factors for treatment failure of patients with ESBL-E bacteremia

In univariate analyses, we evaluated treatment failure according to age, sex, medical history, prior hospitalization, prior antimicrobial therapy, admission to the intensive care unit, PBS, type of antimicrobial therapy, PIP/TAZ minimum inhibition concentration (MIC) ≥ 32 μg/mL, extra-urinary tract infection, and species of pathogen (Table 2). Treatment failure differed significantly according to prior antimicrobial therapy within 30 days, extra-urinary tract infection, and Klebsiella pneumoniae. In a multivariate regression analysis, treatment failure differed significantly according to prior antimicrobial therapy within 30 days and extra-urinary tract infection. However, treatment failure did not show significant differences according to the other variables, including non-carbapenem antimicrobials.

Table 2.

Risk factors for treatment failure in patients with ESBL-Enterobacteriaceae bacteremia

DISCUSSION

The increasing incidence of ESBL-E is a challenge in healthcare, and effective treatment strategies are needed to mitigate the clinical and economic burden. We found a 30-day mortality rate of 16.1%, which is higher than the 10.3% in a 2020 South Korean study [6]. Despite this higher mortality rate, our findings challenge the prevailing preference for carbapenems, for ESBL-E bacteremia.

There was no significant difference in the treatment failure and 30-day mortality rates between the CG and NCG. Extra-urinary tract infection and prior antimicrobial therapy within 30 days were risk factors for treatment failure in patients with ESBL-E bacteremia. This contrasts with the outcomes of the MERINO trial, which compared PIP/TAZ with meropenem. In a randomized clinical trial involving 391 patients with bloodstream infections caused by ESBL-EC and Klebsiella pneumoniae, the primary outcome of 30-day mortality occurred in 12% and 4% of patients receiving PIP/TAZ and meropenem, respectively [12]. PIP/TAZ is not recommended for the treatment of infections other than those of the urinary tract, caused by ESBL-E, even if susceptibility to PIP/TAZ is demonstrated [7]. The choice of antimicrobial therapy is important in patient outcomes, and non-carbapenem alternatives should be considered. However, several design factors in the MERINO trial, such as the short duration of the intervention (median 6 days, compared to a median 13 days of treatment for bacteremia) and and overlap of drug exposure between groups, may have influenced the outcomes and generalizability [12]. Additionally, blaOXA-1 genes were highly prevalent (67%) and phenotypic ESBL production was confirmed in 86% of the isolates [12]. Testing for PIP/TAZ MIC may be inaccurate in the presence of other β-lactamases such as OXA-1, making it unclear if an isolate that tests susceptible to this agent is indeed susceptible. The high prevalence of blaOXA-1 genes suggests an explanation for the high failure rate of PIP/TAZ.

The population of this study was that of patients with ESBL-E bacteremia, whereas the MERINO trial focused on patients with ceftriaxone-resistant EC or Klebsiella spp. This could explain the differences in outcomes and the applicability of results to broader populations. Furthermore, our study highlights the importance of evaluating the characteristics of the individual patient when making treatment decisions. Despite a higher proportion of male patients in the CG, the clinical outcomes were similar in the two groups. This suggests that factors other than the antimicrobial used influence treatment efficacy and patient outcomes. The longer hospitalization duration in the CG raises questions about the cost-effectiveness and effect on patients of prolonged hospitalization. Optimizing hospital stays could improve patient outcomes and reduce healthcare costs, and so further research is needed. Extra-urinary tract infection and prior antimicrobial therapy within 30 days were risk factors for treatment failure in patients with ESBL-E bacteremia. Use of two or more antimicrobials, extended-infusion methods, and adequate drainage of the infection source should be considered for such patients. However, treatment failure did not differ significantly according to age group (≥ 80 years), sex, malignant cancer, prior hospitalization, nosocomial infection, non-carbapenem group, PIP/TAZ MIC ≥ 32 μg/mL, and strain of pathogen.

Our findings challenge the perceived effect of β-lactam/ β-lactamase inhibitors on susceptibility to antimicrobials. Based on data for patients with PIP/TAZ MICs ≤ 16 μg/mL, experts recommend carbapenem therapy for ESBL-E bloodstream infection (BSI) (11% PIP/TAZ vs. 4% carbapenems) and other common sites of infection [13]. In the NCG, PIP/TAZ administration according to MICs against PIP/TAZ demonstrated comparable efficacy in ESBL-E strains, suggesting that broad-spectrum carbapenems may not be necessary in all cases (HR, 1.40; 95% CI, 0.80-9.70 for MICs ≤ 16 μg/mL; HR, 1.90; 95% CI, 0.23-15.60 for MIC = 32 μg/mL; HR, 4.89; 95% CI, 0.38-30.50 for MICs ≥ 64 μg/mL). Although there were few cases with PIP/TAZ MICs ≥ 32 μg/mL (nine patients; median PBS, 0.0; interquartile range, 0.0-2.0), if the disease is not severe and the clinical course is stable, non-carbapenem drugs could be viable options. Furthermore, there was no significant association in the primary outcome among patients receiving antimicrobials other than PIP/TAZ compared to the CG (HR, 0.91; 95% CI, 0.31-2.60; p= 0.857).

We did not investigate the genetic characteristics of the isolates, which may have implications for understanding treatment responses and microbial-resistance patterns. In South Korea, between 2006 and 2011, the most common ESBL types were CTX-M-14 and CTX-M-15 (CTX-M-14, 41.7%; CTX-M-15, 35.0%; CTX-M-24, 8.7%; CTX-M-27, 7.8%; CTX-M-3, 3.9%; CTX-M-57, 2.9%) [14]. Between 2016 and 2017, CTX-M-15 and CTX-M-14 were the most common types [15]. Among the 249 cefotaxime-resistant isolates in a 2019 South Korean multicenter study, only CTX-M-14 and CTX-M-15 were identified [16]. Further molecular epidemiologic studies are needed to guide the management of community-onset infections caused by ESBL-E. Although the prevalence of ESBL enzymes such as OXA-1 have not increased in South Korea, continuous monitoring of other β-lactamases is important in the selection of non-carbapenem drugs.

An ongoing randomized controlled trial is evaluating PIP/TAZ for ESBL-E BSI. In an observational study, there was no difference in the 30-day mortality rate between treatment groups (94% vs. 97%, respectively) [8], and there was no significant difference in the 30-day mortality rate between PIP/TAZ and ertapenem for pyelonephritis or complicated urinary tract infection [17]. This finding mirrors the new IDSA guidelines, which recommend carbapenems for ESBL-E [9]. Antimicrobial drugs should be selected based on the resistance status of the pathogen, and the local antibiotic effectiveness profile.

This study has several limitations. First, it was a single-center retrospective study, and the small sample size could have reduced the statistical power. Second, PBS is used in infectious disease research as an index of the severity of acute disease. A PBS of ≥ 4 is used to define critical illness and an increased risk of mortality [18]. In this study, the disease severity was mild. Further research is needed to determine the effectiveness of non-carbapenem drugs in patients with severe disease. In addition, there were few cases with PIP/TAZ MICs ≥ 32 μg/mL. Moreover, the dynamic landscape of antimicrobial resistance necessitates ongoing evaluation of treatment strategies.

In conclusion, the findings suggest that non-carbapenem drugs can be alternatives for the treatment of ESBL-E bacteremia. This has implications for antimicrobial management and the fight against antibiotic resistance. Such an approach could reduce the selective pressure promoting carbapenem resistance and preserve the effectiveness of these crucial antibiotics. Further large prospective studies are needed to validate the results and refine treatment guidelines.

Article information Continued