Is the use of piperacillin/tazobactam for bacteremia caused by ceftriaxone-nonsusceptible Escherichia coli and Klebsiella pneumoniae an obsolete practice?
Harris and colleagues recently reported the findings of the MERINO trial, which was designed to answer an important question that many clinicians struggle with: Can piperacillin/tazobactam therapy be used as a carbapenem-sparing option to treat ceftriaxone-nonsusceptible Escherichia coli and Klebsiella pneumoniae bloodstream infections (BSIs)?1
Previous investigations into this question yielded conflicting results. A multinational, retrospective cohort found no difference in 30-day mortality between patients treated with β-lactam/β-lactamase inhibitor combinations versus carbapenems for BSIs caused by extended-spectrum β-lactamases (ESBL)—producing Enterobacteriaceae.2 Ng and colleagues found similar outcomes and demonstrated that patients in the piperacillin/tazobactam group were at a lower 30-day risk of acquiring an infection caused by a multidrug-resistant organism compared with the group receiving the carbapenem.3
In contrast, Tamma et al found the adjusted risk of death to be 1.92 times higher when patients received empiric piperacillin/tazobactam compared with a carbapenem in a retrospective review of BSI caused by ESBL-expressing organisms.4
The MERINO trial is the first randomized controlled trial to evaluate piperacillin/tazobactam versus meropenem for the definitive treatment of BSIs caused by ceftriaxone not susceptible E coli or K pneumoniae that remained susceptible to piperacillin/tazobactam by current breakpoints. Patients from 26 sites in 9 countries were stratified over a 3-year period (February 2014 through July 2017) according to the infecting species, presumed source of infection, and severity of illness.
Patients were randomly assigned 1:1 to receive either meropenem 1 g intravenously (IV) every 8 hours or piperacillin/tazobactam 4.5 g IV every 6 hours, with each dose given over 30 minutes. The study was designed to be pragmatic, allowing for the treating physician to change medication after a minimum of 4 days and up to 14 days. The treatment duration was also at the provider’s discretion. The primary efficacy outcome was 30-day all-cause mortality after randomization, with a 5% noninferiority margin. After enrollment of 391 patients, the study was terminated early on the grounds of harm and futility, as assessed by safety monitoring board review.
A total of 23 of 187 patients (12.3%) randomized to piperacillin/tazobactam met the primary outcome compared with 7 of 191 (3.7%) in the meropenem group. When treating with piperacillin/tazobactam, this correlates to a number needed to harm of 12 patients. After subgroup analysis, none of the subgroups met the noninferiority margin, and all favored meropenem. The secondary outcome of clinical and microbiologic success at day 4 of therapy was consistent with the primary outcome, demonstrating again in favor of meropenem (74.6% vs. 68.4%).
Of interest, patients in the piperacillin/tazobactam group had a higher incidence of secondary infections with a multidrug-resistant organism or Clostridium difficile compared with patients who received meropenem (8% vs 4.2%, respectively), but the overall incidence was low.
Phenotypic ESBL production was confirmed in 86% of isolates, and ESBL genes were identified in 85.3% of 293 isolates that underwent whole genome sequencing. A further 10.2% of isolates harbored an acquired plasmid-mediated β-lactamase (AmpC) gene, and 2% harbored both types of genes. Of the ESBL genes, 83.5% were confirmed to have the CTX-M gene, one of the most common ESBL genes in the United States. These data show that using ceftriaxone resistance as a surrogate marker is highly accurate in predicting the presence of an ESBL-producing organism and, more important, the potential outcome of piperacillin/tazobactam therapy.
Furthermore, for institutions that perform ESBL testing by phenotypic or molecular methods, there is a risk of missing AmpC producers and, in turn, greater risk of a poor outcome in this patient population. Evaluation of CTX-M alone by genotypic methods would be just 74% sensitive for the prediction of ceftriaxone-resistant isolates in the population studied.
The authors attempted to investigate the role of the piperacillin/tazobactam minimum inhibitory concentration (MIC, measured by the site’s standard of care method) on mortality but were unable to draw a clear MIC-response correlation. The median MIC of piperacillin/tazobactam was reported to be 2 mcg/mL.
When the researchers further evaluated a small number of available isolates using Etest, a decrease in survival was observed with MIC values >4 mcg/mL, although this observation was not statistically significant. These data bring into question the validity of current Clinical and Laboratory Standards Institute and European Committee on Antimicrobial Susceptibility Testing breakpoints for piperacillin/tazobactam and Enterobacteriaceae, where susceptibility is defined as ≤16 and ≤8 mcg/mL, respectively. The investigators plan to further evaluate the study isolates by reference broth microdilution, which may draw a more defined correlation between MIC and outcome.
The MERINO trial sought to prove that piperacillin/tazobactam could serve as a carbapenem-sparing strategy for ESBL-producing BSIs and instead found the opposite. This is a landmark study in helping to define piperacillin/tazobactam’s role in treating these infections.
Questions that build on these findings merit future study, such as if the use of extended infusion piperacillin/tazobactam dosing would have led to the same outcomes. The BLING-3 study will attempt to further answer this question. This multinational, randomized, controlled trial will compare continuous-infusion meropenem and piperacillin/tazobactam with traditional 30-minute infusions in critically ill patients with sepsis.5
Finally, a question remains as to whether empirical treatment with piperacillin/tazobactam remains appropriate for certain subsets of patients with ceftriaxone nonsusceptible E coli or K pneumoniae BSIs. Using carbapenem therapy for patients traditionally treated with piperacillin/tazobactam (eg, those with a urinary source of infection) seriously compromises antimicrobial stewardship efforts and is associated with significant risk of increasing carbapenem resistance in the community. The MERINO trial data highlight the need for new carbapenem-sparing or carbapenem-minimizing strategies in these populations and also demonstrate the risk associated with the use of piperacillin/tazobactam.
One possible strategy is the use of rapid methods that determine ceftriaxone susceptibility results within the first 2 days of antimicrobial therapy. Such testing may help minimize unnecessary carbapenem doses in patients with BSIs caused by E coli and K pneumoniae. Such technologies exist, including genotypic methods that detect CTX-M (which, based on the MERINO trial data, may be associated with 74% sensitivity) or phenotypic MIC or disk diffusion methods performed directly from blood culture. Incorporating these strategies, either as an escalation or de-escalation method, may prove to have a positive impact on patient outcomes.