Testing for ESBL: Do We Have It All Wrong?

Daniel J. Livorsi, MD, MSc


July 14, 2015

Editorial Collaboration

Medscape &

Carbapenem Therapy Is Associated With Improved Survival Compared With Piperacillin-Tazobactam for Patients With Extended-Spectrum Beta-Lactamase Bacteremia

Tamma PD, Han JH, Rock C, et al
Clin Infect Dis. 2015;60:1319-1325

Routine Testing for Extended-Spectrum Beta-Lactamase

The recent study by Tamma and colleagues and another from 2012[1] provide insight on whether beta-lactam/beta-lactamase-inhibitor combination antibiotics (BLBLIs) have a role in treating infections caused by extended-spectrum beta-lactamase (ESBL) producers.

In 2010, the Clinical & Laboratory Standards Institute lowered the breakpoints of third-generation cephalosporins to improve the detection of ESBLs and AmpC beta-lactamases. As a result, microbiology laboratories are no longer required to test for the presence of ESBLs in Enterobacteriaceae. Proponents of this "new" approach argue that when BLBLIs are active against ESBL producers in vitro, they should also be active in vivo. The presence of an ESBL is therefore less important than the actual minimum inhibitory concentration (MIC).

Opponents argue that clinicians need to know when an ESBL is present, because carbapenems would be the preferred agents. They point to the limited clinical data on using BLBLIs for ESBL infections, the potential presence of undetected resistance mechanisms, and the possibility of an inoculum effect when BLBLIs are used. Who is correct?

To help answer this question, Tamma and colleagues performed a retrospective cohort study at Johns Hopkins University Hospital. Their goal was to determine whether empiric piperacillin-tazobactam was as efficacious as empiric carbapenems in patients with ESBL bacteremia, all of whom received definitive carbapenem therapy. A hospital laboratory database was used to identify 213 eligible cases. The major finding was that, after adjusting for other covariates, the risk for death at 14 days was 1.92 times higher in the patients who received empiric piperacillin-tazobactam (95% confidence interval, 1.07-3.45).


Although it is fair to question the contribution of yet another observational study on this topic, Tamma and colleagues' approach was rigorous. They used multivariable analysis and propensity scoring to minimize confounding. Furthermore, because all patients received a carbapenem once the bloodstream pathogen was identified as an ESBL producer, the point of comparison was truly the effect of piperacillin-tazobactam vs carbapenem as empiric therapy.

A very similar, well-designed, observational study from Spain by Rodriguez-Baño and colleagues[1] evaluated the use of BLBLIs for ESBL-Escherichia coli bacteremias and came to the opposite conclusion. Rodriguez-Baño and colleagues found that if amoxicillin-clavulanic acid and piperacillin-tazobactam are active in vitro, they are acceptable alternatives to carbapenems for ESBL-E colibloodstream infections.[1]

The differences between the studies by Rodriguez-Baño and colleagues and Tamma and colleagues are informative and help to explain their conflicting conclusions. In the study of Rodriguez-Baño and colleagues, > 90% of patients who received piperacillin-tazobactam were administered a dose of 4.5 g intravenously every 6 hours. In contrast, only 39% of the patients receiving piperacillin-tazobactam in the study of Tamma and colleagues were given such a high dose. In the Spanish cohort, about half of the isolates had a piperacillin-tazobactam MIC ≤ 2 µg/mL, which was associated with improved 30-day mortality.[2]

In the Johns Hopkins cohort, 99% of isolates had a piperacillin-tazobactam MIC ≥ 4 µg/mL. Rodriguez-Baño and colleagues found a higher proportion of urinary and biliary infections: 72% vs 26% found by Tamma and colleagues. Finally, Rodriguez-Baño and colleagues exclusively included E coli isolates, whereas Tamma and colleagues also included Klebsiella species and Proteus mirabilis.

In a thoughtful review on this topic, Harris and colleagues[3] argue that BLBLIs may be a reasonable alternative to carbapenems for definitive treatment of less serious ESBL infections if the MIC is low and source control has been achieved. Higher doses of piperacillin-tazobactam are probably warranted, particularly for more serious infections.

An ongoing multicenter, randomized trial is comparing meropenem with piperacillin-tazobactam for ceftriaxone nonsusceptible E coli and K pneumoniae infections. After years of debate, a definitive answer on the need for routine ESBL testing may be just around the corner.



Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as: