Extended-spectrum β-lactamase (ESBL)-producing Gram negative bacilli (GNB) have been classified by the CDC as a serious threat in its report Antibiotic Resistance Threats in the United States, 2013.  Based on conservative estimates, ESBL-producing GNB are responsible for at least 26,000 infections and 1,700 deaths annually.

While carbapenems (e.g., meropenem, imipenem-cilastatin, doripenem, and ertapenem) are considered the drugs of choice for treatment of serious infections such as bacteremia due to ESBL-producing GNB, a rise in carbapenem resistance has led some to investigate carbapenem-sparing regimens to conserve their use and potentially slow further increase in resistance.

Beta-lactamase inhibitors such as tazobactam and clavulanate are active against ESBL in vitro.  As such, beta-lactam/beta-lactamase inhibitor combinations (BL-BLI), such as piperacillin-tazobactam (PTZ), have been explored as carbapenem-sparing options for treatment of ESBL infections.  However, observational studies have yielded mixed results when BL-BLI were compared to carbapenems.  For instance, a post-hoc analysis of 6 observational cohort studies found that 30-day all-cause mortality and length of stay were not statistically different between BL-BLI and carbapenems as empiric or definitive therapy for bloodstream infections (BSI) due to ESBL E coli.  (Rodriguez-Baño J, et al. CID 2012;54:167-74).  Whereas another retrospective study found that the use of PTZ was associated with a 1.9 times higher risk of death as empiric therapy for BSI due to ESBL producers when compared to carbapenems (Tamma PD, et al. CID 2015;60:1319-25).  Data from animal and in vitro models also cast doubts in the efficacy of PTZ against ESBL producers in infections with high bacterial inoculums.

The recently published MERINO trial (Harris PNA, et al. JAMA 2018;320:984-94) was specifically designed to address the controversy of whether BL-BLI produce comparable outcomes to carbapenem in BSI due to ESBL-producing GNB.  This multicenter non-inferiority trial randomized patients to receive PTZ (4.5g q6h as 30-minute infusion) or meropenem (MERO; 1g q8h as 30-minute infusion) as definitive therapy for ESBL-producing E coli or K pneumoniae BSI.  Treatment duration ranged from 4 to 14 days based on the discretion of the treating physician.

Study investigators originally intended to enroll a total of 454 patients but the trial was terminated early after 391 patients were randomized.  The study data safety monitoring board terminated the trial after the 3^rd interim analysis due to harm and futility reasons.

The study groups were well-matched in general.  Overall, the majority of bacteremic episodes were due to E coli (86.2%) with urinary tract being the primary source of bacteremia (54.8%) and study patients had relatively mild severity of illness (92% with Pitt bacteremia score <4).  Notable differences in baseline characteristics included the MERO group had higher mean APACHE II scores (21.0 vs. 17.9%) while the PTZ group had shorter time to receiving microbiologically active antibiotics from onset of infection (5.5 hours vs. 9.6 hours).  By the day of randomization, 41.2% in the PTZ group and 40.3% in the MERO group had resolution of objective signs and symptoms of infections.

The primary outcome of 30-day all-cause mortality was met in 12.3% (23 of 187 patients) in the PTZ group and 3.7% (7 of 191 patients) in the MERO group [risk difference, 8.6% (1-sided 97.5% confidence interval, -∞ to 14.5%); p = 0.90 for noninferiority].  Because the upper bound of the 97.5% confidence interval (14.5%) crossed the pre-determined non-inferiority margin of 5%, PTZ is not non-inferior to MERO for the treatment of BSI due to ESBL E coli or K pneumoniae.  PTZ is likely inferior to MERO when it comes to treatment of these infections (which was presented in 2018 ECCMID in Madrid as abstract O1121).

The secondary outcomes of clinical and microbiological resolution by day 4 after randomization were not significantly different between groups.  Similarly, the rate of microbiological relapse was not different between the PTZ and MERO groups.  Of note, subsequent infection due to carbapenem-resistant organisms was not lower in the PTZ group (3.2%) compared to the MERO group (2.1%), although the follow-up period was short (up to day 30 after randomization) and the number of carbapenem-resistant infections were low.

There are several important limitations to the study. First, ~20% of patients in both group received carbapenems (e.g., ertapenem) as step-down therapy.  The crossover in therapy from PTZ to a carbapenem would tend to bias the primary outcome in favor of PTZ but this was not shown in the final analysis.  Second, only 2 patients from North America (i.e., Canada) were enrolled.  Therefore, the results from this trial may not be generalizable to the patients in the United States.  Third, administration of PTZ and MERO occurred over 30 minutes and results from this study may not be applicable to hospitals that used extended infusion or PK-PD optimized dosing regimens.  Fourth, susceptibility testing were performed using MIC test strips and interpreted based on standards from EUCAST.  This method of susceptibility testing is in contrast to the automated method used by most clinical laboratory.  Interpretative standards were also different between EUCAST and those put forth by CLSI.  Fifth, most deaths in the study cannot be directly attributed to the BSI.  The study investigators defended the use of all-cause mortality as an endpoint and their study conclusion with this and this tweets.  Finally, this trial does not answer the question of whether PTZ can serve as a carbapenem-sparing option if it is not a BSI.

What are some practical implications of the trial for antimicrobial stewards?  Below are some suggestions:

If patient was empirically started on PTZ, consider escalating therapy to a carbapenem if blood culture became positive for ESBL producing GNB, especially if patient is not clinically improving

If 30-minute infusion of PTZ is used in your institution, consider using a carbapenem upfront for a septic patient with a history of ESBL infections

If PTZ is chosen as definitive therapy for BSI due to ESBL organisms, closely monitor clinical status and have a low threshold to escalate to a carbapenem

If PTZ is reported as susceptible despite detecting ESBL in an Enterobacteriaceae at your institution, consider a discussion with microbiology experts and the antimicrobial stewardship team about revising this reporting strategy

The PETERPEN trial which is set to begin enrollment in January 2019 will also compare 30-day all-cause mortality between piperacillin-tazobactam and meropenem for treatment of BSI due to ESBL producers (ClinicalTrials.gov Identified: NCT03671967).  The results from this randomized, controlled, non-inferiority trial will hopefully confirm the findings from the MERINO trial.

To read more about the MERINO trial, please go to https://jamanetwork.com/journals/jama/fullarticle/2702145.

The accompanying editorial can be found here https://jamanetwork.com/journals/jama/fullarticle/2702126.

Written by Phil Chung, PharmD, MS, BCPS