MERINO-2 Trial: Piperacillin-Tazobactam for AmpC Infections

The MERINO-2 trial was a pilot randomized controlled trial that compared piperacillin-tazobactam against meropenem for treating bloodstream infections caused by bacteria that carry chromosomal AmpC enzymes, a group of bugs notoriously tricky to treat. Published in 2021, the trial was too small to give a definitive answer, but it raised a red flag: patients who received piperacillin-tazobactam had more composite treatment failures than those who received meropenem, and the pattern of where things went wrong hinted that the two drugs are not simply interchangeable for these infections.

Why This Trial Existed

MERINO-2 grew directly out of the original MERINO trial, which tested the same drug matchup but for a different set of bacteria. In that earlier study, patients with bloodstream infections caused by ceftriaxone-resistant E. coli or Klebsiella pneumoniae were randomized to piperacillin-tazobactam or meropenem. The trial was stopped early: about 12% of patients in the piperacillin-tazobactam group died within 30 days, compared with roughly 4% in the meropenem group, a threefold difference that made noninferiority impossible to establish.¹ That result cemented carbapenems as the go-to drugs for bloodstream infections caused by extended-spectrum beta-lactamase (ESBL) producers.

But AmpC-producing organisms are a different beast from ESBL producers. Bacteria like Enterobacter species, Serratia marcescens, Citrobacter freundii, Morganella morganii, and Providencia species carry a chromosomal AmpC gene whose expression can ramp up under antibiotic pressure. The clinical question was whether the catastrophic performance of piperacillin-tazobactam against ESBL producers would also apply to these AmpC-carrying species, or whether piperacillin-tazobactam might be safe enough to use as a carbapenem-sparing option.

What MERINO-2 Actually Tested

The trial enrolled adults across multiple countries who had at least one blood culture positive for one of the AmpC-carrying species listed above, provided the isolate tested susceptible to third-generation cephalosporins, piperacillin-tazobactam, and meropenem. The requirement for cephalosporin susceptibility was added partway through the study, after the original MERINO results made it clear that enrolling patients whose bugs might also harbor ESBLs would muddy the picture.² Patients had to be randomized within 72 hours of the initial positive blood culture. The study excluded people with polymicrobial infections, central nervous system sources, or concurrent use of other antibiotics with gram-negative coverage in the first four days.

The primary outcome was a composite measured at 30 days that bundled together mortality, ongoing signs of infection (fever or elevated white blood cell count), microbiological failure (persistence of the same organism in blood cultures), and microbiological relapse (recurrence after initial clearance). This composite was meant to capture the full range of ways treatment can go wrong, not just death.

The Numbers and What They Suggest

In the end, only 72 patients were randomized, a fraction of what a definitive trial would need. Among the 38 patients assigned to piperacillin-tazobactam, 11 (about 29%) met the composite primary outcome, compared with 7 of 34 patients (about 21%) in the meropenem group. That 8-percentage-point difference had wide confidence intervals that crossed zero, meaning the trial could not statistically distinguish the two drugs.³

The more telling detail lay in the subcomponents. Five of 38 patients on piperacillin-tazobactam experienced microbiological failure, meaning their blood cultures never cleared, compared with zero of 34 patients on meropenem. That difference was statistically significant. On the other hand, microbiological relapse ran in the opposite direction: none in the piperacillin-tazobactam arm versus about 9% in the meropenem arm.⁴ In plain terms, piperacillin-tazobactam had a harder time clearing the infection in the first place, but once it did, the bugs were less likely to bounce back. Meropenem cleared infections reliably but saw a few come back. The investigators themselves acknowledged that the trial was too small to draw firm conclusions and described it as generating preliminary data for a larger, definitive study.⁵

The AmpC Problem in Context

Understanding why MERINO-2 matters requires grasping what makes AmpC-producing bacteria different from ordinary gram-negative infections. These species carry a chromosomal gene for AmpC beta-lactamase, an enzyme that can chew up certain antibiotics. Under normal conditions, the gene is expressed at low levels and the bacteria test susceptible in the lab. But exposure to beta-lactam antibiotics can “de-repress” that gene, ramping up enzyme production during treatment. When that happens, drugs like third-generation cephalosporins and piperacillin-tazobactam can lose their effectiveness mid-course, a phenomenon that makes the initial lab susceptibility report misleading.⁶

Not all AmpC-carrying species are equally dangerous in this regard. Enterobacter cloacae is the most notorious offender: deleting a single regulatory gene in the lab produces high-level cephalosporin resistance. Serratia marcescens, by contrast, needs mutations in two separate genes before AmpC is meaningfully de-repressed, and even then the resulting resistance levels are far lower. The AmpC enzyme itself also works less efficiently in Serratia against cephalosporins than the Enterobacter version does.⁷ This variation across species is one reason the field struggles with blanket treatment recommendations: lumping all AmpC producers together oversimplifies the biology.

Where Piperacillin-Tazobactam Stands After Pooled Data

MERINO-2 is one piece of a larger puzzle assembled from observational studies and meta-analyses. A systematic review and meta-analysis comparing piperacillin-tazobactam with cefepime or carbapenems for bloodstream infections caused by chromosomal AmpC producers found no significant difference in mortality between the groups. However, patients treated with piperacillin-tazobactam had roughly 80% higher odds of microbiological failure and about 50% higher odds of clinical failure compared with those treated with cefepime or carbapenems.⁸ The pattern is consistent with what MERINO-2 showed: piperacillin-tazobactam does not seem to get more people killed, but it does a worse job of clearing the bacteria.

That distinction between mortality and microbiological clearance is clinically meaningful. A patient whose bloodstream infection lingers stays sicker longer, may need additional courses of antibiotics, and accumulates more days in the hospital. Even if the infection does not ultimately prove fatal, persistent bacteremia drives complications.

Cefepime as a Carbapenem-Sparing Alternative

If piperacillin-tazobactam is a shaky carbapenem substitute for AmpC-producing infections, what about cefepime? Cefepime is a fourth-generation cephalosporin that is inherently more resistant to hydrolysis by AmpC enzymes than third-generation cephalosporins are. A meta-analysis pooling seven studies found no statistically significant difference in mortality between cefepime and carbapenems for invasive infections caused by inducible AmpC-producing bacteria.⁹ A separate multicenter study specifically looking at high-dose cefepime for bacteremia caused by AmpC producers found a similar result: no increased 30-day mortality compared with carbapenem therapy.¹⁰

Cefepime has its own downsides, though. At high serum levels it can cause neurotoxicity, with symptoms ranging from confusion to seizures, particularly in patients with impaired kidney function. That risk has prompted work on optimized dosing strategies. A real-world validation of a reduced-dose continuous infusion cefepime regimen found that over 95% of patients hit the desired pharmacodynamic target while maintaining high clinical efficacy.¹¹ Continuous or extended infusions keep drug levels more stable and may help thread the needle between efficacy and toxicity. For many hospitals, this approach makes cefepime a realistic carbapenem-sparing option for AmpC producers, where piperacillin-tazobactam looks less reliable.

Kidney Injury and Drug Safety

An underappreciated wrinkle in choosing between these antibiotics is the risk of acute kidney injury. Piperacillin-tazobactam has been associated with a substantially higher likelihood of kidney injury compared with both cefepime and meropenem. One study using statistical methods to control for baseline patient differences found that piperacillin-tazobactam carried roughly seven times the odds of acute kidney injury compared to cefepime, and a similar increase compared to meropenem.¹² This risk is especially relevant in critically ill patients who already have compromised kidney function. It adds another reason to think twice before defaulting to piperacillin-tazobactam for serious infections, even when the lab report says the bug is susceptible.

Detecting AmpC Production in the Lab

One reason clinicians sometimes end up using piperacillin-tazobactam for AmpC-producing infections is that the lab report may not flag the AmpC issue. Standard susceptibility testing tells you whether the bacteria are susceptible at the moment of testing, but it does not reliably predict whether AmpC will be de-repressed during treatment. Detecting AmpC production requires additional phenotypic or genotypic tests that many routine clinical labs do not perform.

The available phenotypic tests vary widely in accuracy. A head-to-head evaluation of six methods for detecting AmpC found that the best-performing commercial test achieved roughly 95% accuracy, while basic cefoxitin screening was much less sensitive, catching only about 75% of AmpC producers, with a specificity as low as 25%.¹³ An earlier practical approach using cefoxitin as an initial screen followed by a confirmatory double-disk synergy method showed better sensitivity, exceeding 97%, though this two-step process takes longer.¹⁴ Molecular methods like PCR are the gold standard but are not universally available. The upshot is that if a patient has a bloodstream infection caused by an Enterobacter or one of the other usual AmpC suspects, the clinician often has to make a treatment decision based on species identification alone, without confirmation of AmpC status.

The Carbapenem Stewardship Dilemma

The whole reason trials like MERINO-2 exist is the tension between two imperatives. On one hand, carbapenems work. They are reliably effective against AmpC producers because AmpC enzymes generally cannot break them down. On the other hand, every dose of carbapenem used is a dose that selects for carbapenem-resistant organisms, which are among the most dangerous infections in medicine. Stewardship programs across the world have been set up specifically to curb unnecessary carbapenem use, because overuse has contributed to the spread of carbapenem-resistant bacteria.¹⁵

MERINO-2’s small size means clinicians are left making judgment calls. For a patient who is hemodynamically stable with a bloodstream infection caused by a species known to have lower de-repression risk, like Serratia marcescens or Morganella morganii, many infectious disease specialists feel comfortable using cefepime or even considering piperacillin-tazobactam if the isolate tests susceptible. For a critically ill patient with an Enterobacter cloacae bloodstream infection and an undrained abscess, the calculus shifts heavily toward a carbapenem. The severity of illness, the source of infection, and the specific organism all factor in, and no single trial has resolved the debate for every clinical scenario.

Oral Step-Down After Intravenous Therapy

Once a patient with an AmpC-producing bloodstream infection has stabilized on intravenous antibiotics, the question of when and how to switch to oral therapy comes up. A study of patients treated for AmpC-producing bloodstream infections found that those who stepped down to oral antibiotics had shorter hospital stays, with a median of 14 days compared with 18 days for those who stayed on intravenous therapy, without any increase in recurrence.¹⁶

The choice of oral agent matters. A systematic review of step-down therapy for gram-negative bacteremia more broadly found that fluoroquinolones may reduce the chance of infection recurrence compared with oral beta-lactams, likely because fluoroquinolones achieve higher and more reliable blood levels after oral dosing.¹⁷ For AmpC producers specifically, fluoroquinolones have the added advantage of working through a completely different mechanism than beta-lactams, sidestepping the AmpC enzyme altogether. The catch is that fluoroquinolone resistance is increasingly common, and these drugs carry their own side effect profile, including tendon problems and nerve damage. Still, when the isolate tests susceptible and the patient is improving, an early switch to an oral fluoroquinolone can shave days off a hospital stay without obvious harm.

MERINO-3 and What Comes Next

The investigators behind MERINO-2 have continued the program with MERINO-3, a larger trial comparing ceftolozane-tazobactam with meropenem for bloodstream infections caused by ESBL and AmpC producers. Ceftolozane-tazobactam is a newer beta-lactam/inhibitor combination designed to resist hydrolysis by many of the enzymes that defeat older drugs. The rationale is that if a non-carbapenem agent can match meropenem’s outcomes in a well-powered trial, it would provide a strong evidence-based alternative for stewardship programs to lean on.¹⁸

Many of the uncertainties that MERINO-2 left unresolved remain open. These include whether the risk of AmpC de-repression varies meaningfully enough between species to justify tailored antibiotic choices, how to handle high-inoculum infections like ventilator-associated pneumonia where antibiotic penetration and bacterial burden complicate things, and whether de-escalation from a carbapenem once a patient stabilizes is safe in practice.¹⁹ MERINO-2 was always designed as a pilot, and it did exactly what a pilot trial should: it showed that recruiting for this question is feasible, generated enough signal to justify a larger investment, and sharpened the specific questions a definitive trial would need to answer. Until that definitive trial arrives, clinicians will keep balancing the reliable efficacy of carbapenems against the long-term consequences of using them more than necessary.