By Matthew Hitchcock, MD, MPH

Infectious Diseases Fellow, Department of Medicine, Division of Infectious Diseases and Geographic Medicine, Stanford University

Dr. Hitchcock reports no financial relationships relevant to this field of study.

SYNOPSIS: In an open-label, randomized, noninferiority trial evaluating the efficacy of piperacillin-tazobactam vs. meropenem for definitive therapy in treating bacteremia caused by extended-spectrum beta-lactamase-producing Escherichia coli and Klebsiella pneumoniae, piperacillin-tazobactam therapy did not result in noninferior 30-day, all-cause mortality compared to meropenem. Investigators stopped the trial early due to futility.

SOURCE: Harris PNA, Tambyah PA, Lye DC, et al. Effect of piperacillin-tazobactam vs meropenem on 30-day mortality for patients with E. coli or Klebsiella pneumoniae bloodstream infection and ceftriaxone resistance: A randomized clinical trial. JAMA 2018;320:984-994.

With the steady progression of antimicrobial resistance in gram-negative bacteria, it is critical to maintain the effectiveness of broad-spectrum antibiotics. As carbapenems have emerged as a frontline treatment for infections caused by many resistant organisms, including those that produce extended-spectrum beta-lactamases (ESBLs),1 and as ESBL-producing organisms have become more prevalent,2 there has been significant interest in identifying alternative, carbapenem-sparing regimens to reduce the selective pressure favoring the emergence of carbapenem-resistant organisms. One such approach has involved the use of beta-lactam/beta-lactamase inhibitors (BLBLIs), such as piperacillin-tazobactam, as tazobactam can inhibit many ESBLs.3 Prior retrospective studies have evaluated use of BLBLIs vs. carbapenems in both empirical and definitive therapy of infections caused by ESBL-producing organisms, with mixed results and multiple studies showing no difference in outcomes.4-6 However, at least one recent study showed improved outcomes with the use of carbapenems compared to piperacillin-tazobactam.7

Because of this ongoing debate, the MERINO trial was designed to evaluate the hypothesis that piperacillin-tazobactam was noninferior to meropenem for the definitive treatment of bloodstream infection caused by Escherichia coli and Klebsiella pneumoniae isolates resistant to third-generation cephalosporins in a multicenter, open-label, randomized, controlled trial.8 It was designed to mimic clinical practice, allowing for empiric therapy at the discretion of the treating clinician prior to randomization for definitive therapy once microbiological results were available. Patients were eligible for inclusion if they had at least one blood culture positive for these organisms with local susceptibility results showing resistance to ceftriaxone or cefotaxime and susceptibility to both piperacillin-tazobactam and meropenem. Important exclusion criteria included drug allergy to either study drug or drug class, no expectation of survival beyond 96 hours, treatment without curative intent, polymicrobial bacteremia (with an exception for likely skin contaminants), and requirements for additional gram-negative antimicrobial therapy.

Patients were randomized within 72 hours of initial blood culture collection in a 1:1 fashion to meropenem 1 g q8 hours or piperacillin-tazobactam 4.5 g q6 hours intravenously and stratified by study site, infecting species, presumed infection source (urinary tract vs. elsewhere), and disease severity. Treatment was administered for a minimum of four days and a maximum of 14 days, with the option to stop all antibiotics on day 5, continue with the study therapy, or switch to an alternative agent as de-escalation. The primary outcome of the study was all-cause mortality at 30 days. The coordinating laboratory also collected and analyzed organisms isolated from study patients to confirm minimum inhibitory concentrations (MICs) to meropenem and piperacillin-tazobactam, establish ESBL presence by disk testing with clavulanate, and evaluate ESBL genes by whole-genome sequencing. A 5% margin for noninferiority was prespecified.

During the trial period, investigators screened 1,646 patients and enrolled 378 patients. Baseline patient characteristics generally were balanced, although patients in the meropenem arm were more likely to have a urinary source of infection and higher APACHE score at randomization. Patients in the piperacillin-tazobactam arm were more likely to have immune compromise.

The authors noted that 40.3% of all patients, balanced across arms, showed resolution of signs of infection at the time of study entry following receipt of empiric antibiotic therapy. Investigators stopped the trial early when interim analysis showed a significantly higher mortality rate in the piperacillin-tazobactam arm and full enrollment was not expected to demonstrate non-inferiority. Overall, 12.3% (23 of 187) of patients randomized to piperacillin-tazobactam died within 30 days, compared to 3.7% (7 of 191) of patients randomized to meropenem in the intention-to-treat analysis, and results were similar in the per-protocol analysis. Statistical adjustment within subgroups did not affect the primary outcome, as the noninferiority margin was not met and meropenem was favored in all subgroups. Patients received a similar mean duration of study drug (7.3 days for piperacillin-tazobactam and 7.6 days for meropenem) and total duration of antibiotic therapy (13.2 days vs. 13.7 days). Eighteen percent of patients received empiric therapy congruent with study allocation prior to randomization (20.7% with piperacillin-tazobactam and 15.2% with meropenem). Conversely, 13.8% of patients randomized to receive piperacillin-tazobactam were treated empirically with a carbapenem prior to allocation and 26.2% of patients randomized to meropenem were treated empirically with a BLBLI prior to allocation. Notably, 20.2% of patients randomized to piperacillin-tazobactam received a carbapenem (typically ertapenem) as step-down therapy, while only 2.6% of patients randomized to meropenem received a BLBLI as step-down therapy and 20.4% received a different carbapenem. Fifty-six percent of patients received no additional antibiotics after completion of the allocated therapy, with a similar proportion in each arm.

In the analysis of secondary outcomes, clinical and microbiological resolution by day 4 was similar in the treatment groups and the time to resolution of infectious symptoms also was similar. There was no difference in the rates of microbiological relapse, secondary infection with a multidrug-resistant organism, and C. difficile infection. The recovery of carbapenem-resistant organisms as a secondary infection was rare, and rates were similar across the arms (3.2% with piperacillin-tazobactam and 2.1% with meropenem).

In microbiological analysis, 306 total isolates, representing 80% of the participants, were available for additional testing (n = 266 E. coli and n = 40 K. pneumoniae). The median MIC for piperacillin-tazobactam was 2 mg/L (interquartile range [IQR], 1.5 to 4 mg/L); 12 (3.9%) isolates were resistant by EUCAST breakpoint for susceptibility, (≤ 8 mg/L), but only four (1.3%) isolates were resistant by the CLSI breakpoint (≤ 16 mg/L). Investigators found no significant difference in the median piperacillin-tazobactam MICs across the treatment arms and no association between piperacillin-tazobactam MIC and mortality rate. Regarding meropenem, 99.7% of isolates were susceptible, with a median MIC of 0.23 mg/L (IQR, 0.016 to 0.032 mg/L); the single resistant isolate was found to carry a variant OXA-48 carbapenemase. Researchers found confirmation of phenotypic ESBL production in 86.0% of isolates (85.0% of E. coli and 92.5% of K. pneumoniae) and ESBL genes were found in 85.3% of isolates by whole-genome sequencing, which were predominately blaCTX-M-type (83.5%). Ten percent of isolates carried acquired ampC genes and 2.0% carried both an ESBL and ampC.


The MERINO trial was designed to reflect the actual management of infected patients in clinical practice and focused on the question of whether BLBLIs such as piperacillin-tazobactam are an adequate substitution for carbapenems in treating established ESBL infections, and it answered rather definitively in the negative. The increased mortality with BLBLI therapy appears to be independent of the presence of overt phenotypic resistance, as the vast majority of organisms were susceptible to piperacillin-tazobactam, which suggests that standard phenotypic methods may not accurately reflect the effectiveness of BLBLIs in vivo against organisms with complex resistance mechanisms. Although empiric therapy prior to randomization was outside the control of the study — by design — it seems less likely that this would explain the increased mortality with BLBLI therapy. The majority of patients in both arms were treated with an empiric antibiotic from a different class than either study drug, with many crossing over at the time of randomization. Cephalosporins would be expected to have limited efficacy against ESBL-producing gram-negative organisms. In addition, 40% of patients in each arm showed resolution of signs of infection prior to randomization and receipt of any study drug. It also is unlikely that the quality of supportive care adversely affected the primary outcome in either direction, as patients were enrolled in both arms at all study sites, and none of the sites had vastly disparate mortality rates compared to the overall findings. No patients from the United States were enrolled, and while that theoretically may limit the generalizability of these findings to an American context, the global spread of ESBL resistance mechanisms likely abrogates this limitation.

Because of the study design, the question of whether carbapenems are preferred for empiric therapy for patients with a reasonable likelihood of having a bloodstream infection with an ESBL-producing gram-negative bacteria was not addressed. The answer to that question is likely to be highly context-specific and based on the local prevalence of ESBL-producing organisms and the severity of illness. However, the MERINO results suggest that switching to a carbapenem immediately upon identification of an ESBL-producing organism may make up for suboptimal empiric therapy, given that 86% of patients in the meropenem arm were treated empirically with a noncarbapenem antibiotic. This may help balance the need for stewardship of carbapenem use to maintain future efficacy while also ensuring patients receive adequate empiric therapy.

Questions also remain about whether an extended infusion of piperacillin-tazobactam or whether newer BLBLI agents such as ceftazidime-avibactam or ceftolozane-tazobactam may be more effective than piperacillin-tazobactam at standard dosing. Until these questions can be addressed in similar randomized, controlled trials, the results of the MERINO trial indicate that when an ESBL-producing, gram-negative bacteremia is found, carbapenems are preferred over BLBLIs for definitive therapy.


  1. Paterson DL, Ko WC, Von Gottburg A, et al. Antibiotic therapy for Klebsiella pneumoniae bacteremia: Implications of production of extended-spectrum beta-lactamases. Clin Infect Dis 2004;39:31-37.
  2. McDanel J, Schweizer M, Crabb V, et al. Incidence of extended-spectrum β-lactamase (ESBL)-producing Escherichia coli and Klebsiella infections in the United States: A systematic literature review. Infect Control Hosp Epidemiol 2017;38:1209-1215.
  3. Drawz SM and Bonomo RA. Three decades of beta-lactamase inhibitors. Clin Microbiol Rev 2010;23:160-201.
  4. Gutiérrez-Gutiérrez B, Pérez-Galera S, Salamanca E, et al. A multinational, preregistered cohort study of β-lactam/β-lactamase inhibitor combinations for treatment of bloodstream infections due to extended-spectrum-β-lactamase-producing Enterobacteriaceae. Antimicrob Agents Chemother 2016;60:4159-4169.
  5. Ng TM, Khong WX, Harris PN, et al. Empiric piperacillin-tazobactam versus carbapenems in the treatment of bacteraemia due to extended-spectrum beta-lactamase-producing Enterobacteriaceae. PLoS One 2016;11:e0153696.
  6. Muhammed M, Flokas ME, Detsis M, et al. Comparison between carbapenems and β-lactam/β-lactamase inhibitors in the treatment for bloodstream infections caused by extended-spectrum β-lactamase-producing Enterobacteriaceae: A systematic review and meta-analysis. Open Forum Infect Dis 2017;4:ofx099.
  7. Tamma PD, Han JH, Rock C, et al. Carbapenem therapy is associated with improved survival compared with piperacillin-tazobactam for patients with extended-spectrum β-lactamase bacteremia. Clin Infect Dis 2015;60:1319-1325.
  8. Harris PNA, Tambyah PA, Lye DC, et al. Effect of piperacillin-tazobactam vs meropenem on 30-day mortality for patients with E. coli or Klebsiella pneumoniae bloodstream infection and ceftriaxone resistance: A randomized clinical trial. JAMA 2018;320:984-994.