Studies Contradicting the MERINO Trial
Yes, multiple observational studies and systematic analyses contradict the MERINO trial's findings, showing no significant mortality difference between piperacillin-tazobactam and carbapenems for ESBL-producing Enterobacterales infections, particularly in lower-risk scenarios. However, the evidence quality is substantially lower than MERINO's randomized controlled trial design.
The Core Conflict
The 2022 ESCMID guidelines explicitly acknowledge "the conflict between the MERINO RCT and the observational data" 1. This represents the central tension in managing ESBL infections.
Evidence Supporting Piperacillin-Tazobactam
Observational studies consistently show non-inferiority:
Nine retrospective observational studies comparing carbapenems with carbapenem-sparing regimens (mainly piperacillin-tazobactam, quinolones, and aminoglycosides) found no significant outcome difference in mortality or clinical failure 1
A 2025 Swedish multicenter study (644 patients) demonstrated piperacillin-tazobactam was non-inferior to carbapenems for 30-day mortality in ESBL bloodstream infections, with an absolute risk difference of -0.4% (statistically significant non-inferiority with 5% delta limit) 2
Urinary tract infections specifically: Multiple studies show piperacillin-tazobactam is non-inferior for both complicated and uncomplicated UTIs caused by ESBL-producing organisms, with comparable clinical cure, microbiological cure, and clinical response rates 3, 4, 5
Pyelonephritis: A multicenter observational study (186 patients) found no difference in recurrent infection rates (20% vs 25%), symptom resolution, or 30-day mortality between piperacillin-tazobactam and carbapenems 5
Nosocomial pneumonia: A 2022 study (136 patients with ESBL-producing K. pneumoniae) showed comparable 28-day mortality, clinical cure, and microbiological cure when MICs were ≤8 mg/L 6
Critical Nuances in the Evidence
The observational data has significant limitations:
All comparative observational studies were rated as moderate to high risk of bias due to small sample sizes, residual confounding despite adjustment, and baseline differences between groups 1
The overall certainty of evidence for lack of carbapenem advantage was judged as low for low-risk BSI and very low for high-risk BSI 1
Some studies showed odds ratios trending toward carbapenems for mortality, though not reaching statistical significance 1
One recent study supports MERINO's findings:
A 2026 Italian emergency department study (264 patients) found higher in-hospital mortality with empirical piperacillin-tazobactam versus carbapenems (14.3% vs 3.5%, adjusted risk difference 10%) for ESBL bloodstream infections 7
Notably, this difference disappeared when comparing definitive therapy (6% vs 9% mortality), suggesting the issue may be with empirical rather than targeted use 7
Risk Stratification is Critical
The key distinction is infection severity and source:
Low-risk infections (uncomplicated UTI, pyelonephritis without septic shock): Observational data supports piperacillin-tazobactam as reasonable alternative 1, 3, 4, 5
High-risk infections (septic shock, high-risk bloodstream infections): MERINO's findings remain most applicable, favoring carbapenems 1
MIC considerations: When piperacillin-tazobactam MICs are ≤8 mg/L, outcomes appear comparable; higher MICs associated with worse outcomes 6
Common Pitfalls
Applying observational data to high-risk patients: The observational studies showing non-inferiority predominantly included lower-risk patients and urinary sources
Ignoring the quality hierarchy: A single well-designed RCT (MERINO) carries more weight than multiple observational studies with high risk of bias
Overlooking the empirical vs. definitive distinction: The Italian study suggests empirical piperacillin-tazobactam may be problematic even when definitive therapy is acceptable 7
Antibiotic stewardship considerations: The observational evidence exists partly because of the "underlying stewardship consideration for this comparison was the association between carbapenem use and CRE" 1