As the antibiotic pipeline produces new therapies, clinicians must understand each agent’s specific role in manage­ment of patients infected by multidrug-resistant pathogens.
The World Health Organization (WHO) lists the emergence of antibiotic resistance as 1 of the top 10 threats to public health.1 Emergence of resistance is particularly problematic among patients with infections due to gram-negative bacteria (GNB). The terms multidrug resistant (MDR), extensively drug resistant, and pan-drug resistant are often used to char­acterize patterns of drug resistance exhibited by GNB.2 Strains that are resistant to antibiotics in multiple classes pose a substantial clinical challenge, often requiring use of novel or “last resort” agents.3
Fortunately, there has been impressive drug development in recent years in response to growing threats associated with antibiotic-resistant GNB that cause invasive infections in humans, including Pseudomonas aeruginosa, Klebsiella pneumoniae, and Acinetobacter baumannii. Two recent additions to the arma­mentarium for the treatment of adult patients with GNB infections are imipenem/cilastatin/ relebactam (IMI/REL) and cefiderocol. Below, we provide a general overview of each agent, describe its mechanism of action, detail the microbiologic activity against common highly resistant GNB, and highlight major clinical trial findings.
Indication and Dosing
IMI/REL is a combination of imipenem, a carbapenem anti­bacterial; cilastatin, a renal dehydropeptidase inhibitor; and relebactam, a non-β-lactam β-lactamase inhibitor. In July 2019, the US Food and Drug Administration (FDA) approved IMI/ REL (Recarbrio) to treat complicated intra-abdominal infec­tions (cIAIs) and complicated urinary tract infections (cUTIs), including pyelonephritis, caused by susceptible GNB in adult patients who have limited or no alternative treatment options. It is dosed at 1.25 g (imipenem 500 mg, cilastatin 500 mg, relebactam 250 mg) every 6 hours in patients with creatinine clearances 90 mL/min or greater, and dosage adjustments are required in patients with renal impairment.4
Although imipenem is active against many GNB, its activity is markedly enhanced with the addition of relebactam, a bicyclic diazabicyclooctane that is structurally similar to avibactam. Relebactam exhibits a positively charged piperidine ring added to the carbonyl group, which reduces efflux of relebactam from certain GNB.5 It inhibits Ambler class A (eg, extend­ed-spectrum β-lactamases, carbapenemases), and class C ceph­alosporinases, effectively restoring imipenem activity against resistant K pneumoniae carbapenemase (KPC)— and AmpC-producing Enterobacteriaceae.
Like avibactam, relebactam does not inhibit metallo- β-lactamases (MBLs) or most carbapenem-destructive OXA- β-lactamases. Furthermore, porin deficiency in K pneumo­niae appears to play a significant role in its susceptibility, as in vitro resistance has been reported with OmpK36 porin mutations in K pneumoniae.6 The activity of IMI/REL against non-MBL carbapenemase-producing K pneumoniae isolates was best highlighted in an analysis of 314 consecutive clin­ical strains collected across 18 hospitals in Greece between November 2014 and December 2016. Overall, IMI/REL inhib­ited 98.0% of KPC-producing isolates at concentrations of ≤2 mg/L. Reduced activity of IMI/REL was rarely detected (2%) and was associated with chromosomal factors (OmpK35 disruption and/or mutated OmpK36). Against KPC-producing K pneumoniae, relebactam lowered the imipenem minimum inhibitory concentration 90th percentile (MIC90) from >64 to 1 mg/L. However, relebactam provided only weak poten­tiation of imipenem activity against K pneumoniae with class D OXA-48-like enzymes.7
IMI/REL also has potent in vitro activity against MDR P aeruginosa. Unlike vaborbactam, which has a limited propen­sity to improve meropenem susceptibility for P aeruginosa, relebactam appears to substantially potentiate imipenem activity against imipenem-resistant P aeruginosa isolates.8 Neither imipenem nor relebactam appears to be a substrate for the efflux pumps present in P aeruginosa, and relebactam preserves imipenem activity in the setting of AmpC hyperpro­duction.9 The IMI/REL combination reduces imipenem MICs by approximately 8-fold in imipenem-nonsusceptible P aerugi­nosa isolates. When relebactam was added to clinical isolates from the SMART global surveillance program (n = 993, 2009; n = 1702, 2011; n = 5953, 2015; n = 6165, 2016), imipenem susceptibilities increased to 87.6, 86.0, 91.7, and 89.8%, respec­tively, compared with 68.4, 67.7, 70.4, and 67.3%, respectively, for imipenem alone.10
Two phase 2 clinical trials—1 for cUTI and 1 for cIAI—compared IMI/REL with imipenem/cilastatin.11,12 Both studies evaluated 2 dosing regimens for IMI/REL (imipenem 500 mg/ cilastatin 500 mg/relebactam 250 mg and imipenem 500 mg/ cilastatin 500 mg/relebactam 125 mg). The cUTI trial included 298 adult patients, with 99 treated with the now FDA-approved dose of IMI/REL; the cIAI trial included 347 patients, with 117 treated with the FDA-approved dose. In both trials, clinical and microbiologic efficacy was >95% for all treatment groups, demonstrating noninferiority for IMI/REL versus imipenem/ cilastatin. The most common treatment-associated adverse events (AEs) across groups were gastrointestinal related and headache.
To generate clinically applicable data with IMI/REL for the treatment of adult patients with imipenem-nonsusceptible GNB infections, a randomized, controlled, double-blind trial, RESTORE-IMI 1, was conducted.13 This noninferential, descriptive study compared IMI/REL with colistin-based therapy (colistin plus imipenem/cilastatin) for patients with hospital-acquired bacterial pneumonia and ventilator-associated bacterial pneumonia (HABP/VABP), cUTIs, or cIAIs caused by imipenem-nonsusceptible (but colistin- and IMI/REL-susceptible) pathogens.
Among the 31 patients (n = 21, IMI/REL; n = 10, colistin-based therapy) in the modified microbiologic intent-to-treat population (qualifying baseline pathogen and ≥1 dose study treatment), there were 11 HABP/VABP, 16 cUTI, and 4 cIAI cases, and the predominate pathogen was P aeruginosa. Overall, a favorable response was observed in ~70% of patients in each treatment group.* Day 28 favorable clin­ical response was higher among patients who received IMI/ REL versus colistin-based therapy (71% and 40%, respec­tively; 90% CI, 1.3-51.5), and 28-day mortality was lower among patients in the IMI/REL group relative to the colis­tin-based therapy group (10% and 30%, respectively; 90% CI, —46.4 to 6.7).
Notably, there were no instances of treatment-emergent IMI/REL nonsusceptibility. Treatment-emergent nephrotox­icity was significantly less frequent with IMI/REL than with colistin-based therapy (10% and 56%, respectively [95% CI, —69.1% to –18.4%). Although statistical inferences could not be drawn because of the trial’s descriptive nature, the results showed that IMI/REL was effective for treating imipenem-nonsusceptible GNB infections without the neph­rotoxicity associated with colistin.
IMI/REL was compared with piperacillin/tazobactam in RESTORE-IMI 2, a global, multicenter, randomized, nonin­feriority trial of adult patients with HABP/VABP. In a press release, Merck indicated that IMI/REL met both the primary end point (day 28 all-cause mortality) and key secondary end points (eg, clinical response at early follow-up) of statistical noninferiority compared with piperacillin/tazobactam in the modified intent-to-treat population. Rates of AEs observed in the trial were also reported to be similar in both groups. Merck plans to present the full data at a scientific congress in 2020.14
Indication and Dosing
Cefiderocol (Fetroja), a first-in-class siderophore cephalo­sporin,15 is currently FDA approved for the treatment of adult patients with cUTIs, including pyelonephritis, caused by susceptible GNB who have limited or no alternative treatment options. The dosing regimen of cefiderocol is 2 g adminis­tered every 8 hours by intravenous infusion over 3 hours. Adjustments to the dosage and/or frequency of administra­tion are recommended for patients with renal impairment or augmented renal clearance.16
Like all β-lactams, cefiderocol exerts a bactericidal effect by binding to penicillin-binding proteins (PBPs) and inhibiting cell wall synthesis. Cefiderocol also has a distinctive mech­anism for efficiently penetrating the outer membrane of GNB. As a siderophore cephalosporin, it binds to ferric iron, an essential nutrient for bacterial growth and virulence, and is actively transported across the outer membrane common to all GNB into the periplasmic space. This results in high concentrations of cefiderocol in the periplasmic space, where it can then bind to PBPs and inhibit cell wall synthesis. It is also stable against both serine and MBL carbapenemases.17-19
Data from global surveillance studies for cefiderocol demonstrated potent in vitro activity against a wide spectrum of MDR-GNB, including carbapenem-resistant A baumannii, P aeruginosa, Enterobacteriaceae, and Stenotrophomonas maltophilia.20 Of the approximately 30,000 isolates collected in the SIDERO-WT surveillance studies (2014-2017), just 161 had a cefiderocol MIC >4 mg/L. Of the 11,168 isolates from the United States, 99.7% demonstrated a MIC of ≤4 mg/L; 19 showed a cefiderocol MIC >4 mg/L (9 A baumannii, 6 Enterobacteriaceae, 3 Burkholderia cepacia complex, 1 P aeruginosa). Among the 1272 meropenem-nonsusceptible isolates of Enterobacteriaceae, P aeruginosa, and A baumannii collected as part of the 2014 surveillance study, MIC values for cefiderocol were ≤4 mg/L against 97.7% of tested isolates. Notably, susceptibility was demon­strated for 100% of IMP, OXA-58, KPC, VIM, and OXA-48- like-producing isolates; 99.3% of carbapenemase-negative isolates; 97.2% of OXA-23-positive isolates; 95.2% of OXA-24- positive isolates; 91.7% of GES-positive isolates; and 64.3% of NDM-positive isolates. Cefiderocol MICs were ≤4 mg/L against 99.3% of 136 colistin-resistant Enterobacteriaceae, including isolates carrying mcr-1.21
To date, cefiderocol has been evaluated in 3 clinical trials (1 phase 2, cUTI; 1 phase 3, HABP/VAP; 1 open-label, carbapenem-resistant isolates). The phase 2 cUTI trial was a randomized, double-blind, multicenter, noninferiority trial involving adult patients with cUTIs22 at risk of MDR gram-neg­ative infection. Patients were randomly assigned 2:1 to cefid­erocol 2 g infused over 1 hour 3 times daily (n = 303) or imipenem-cilastatin 1 g infused over 1 hour 3 times daily (n = 149) for 7 to 4 days. Of note, the currently recommended cefiderocol infusion duration is 3 hours.
Among the 371 patients (n = 252, cefiderocol group; n = 119, imipenem-cilastatin group) that had qualifying GNB, the primary efficacy end point, a composite of clinical and microbiological response, at test of cure (TOC) was observed more frequently in the cefiderocol group than in the imipenem/cilastatin group (73% vs 55%; adjusted treatment difference, 18.58%; 95% CI, 8.23-28.92; P = .0004). The difference in the primary composite end point at TOC was due to greater microbiologic eradication in the cefiderocol arm. Both drugs were well tolerated, and the AE profile of cefiderocol was similar to that of imipenem/ cilastatin, with the most commonly reported effects being gastrointestinal.
The CREDIBLE-CR study, a descriptive, open-label trial, investigated the efficacy and safety of cefiderocol versus best available therapy (BAT) in patients with evidence of gram-negative, carbapenem-resistant (CR) pathogens. Patients were randomized 2:1 to cefiderocol or BAT. Selection of BAT in the control arm was at the discretion of the treating clinician; most patients received colistin-based therapy. Cefiderocol was dosed at 2 g every 8 hours administered as a 3-hour infusion. In total, 150 patients were enrolled. The primary end point population (CR microbiologic intent-to-treat [MITT]) included 80 patients who received cefiderocol and 38 patients who received BAT.
The 3 most commonly reported pathogens in this study were CR A baumannii, CR K pneumoniae, and CR P aeruginosa. Most patients had HABP, VABP, or health care—associated pneumonia (HCAP) (45%), followed by bloodstream infections (BSIs) and/or sepsis (31%) and cUTIs (24%). Of note, there were several baseline imbalances between the cefiderocol and BAT groups. For example, the cefiderocol group had a greater proportion of patients ≥65 years old, and all 5 patients with S maltophilia were randomized to cefiderocol. In contrast, although the rate was high in both groups, a higher percentage of patients in the BAT group had prior antibacterial therapy in the 2 weeks before randomization than in the cefiderocol group (100% vs 92.1%, respectively).
In the CR MITT population, clinical cure rates for all infection sites combined in patients treated with cefid­erocol compared with those treated with BAT at TOC were 52.5% versus 50.0%, respectively, and clinical failure rates at TOC were 33.8% versus 36.8%, respectively. Likewise, microbiologic eradication rates for cefiderocol versus BAT at TOC were 31.3% and 23.7%, respectively, and microbiologic failure rates at TOC were 20.0% and 26.3%, respectively.
There was a higher all-cause mortality rate in cefiderocol-treated patients compared with BAT-treated patients at days 14 and 28 and at end of study (18.8% vs 12.2%, 24.8% vs 18.4%, and 33.7% vs 18.4%, respectively). At day 14, day 28, and end of study, all-cause mortality was higher in patients with baseline HAP/VAP/HCAP and BSI/sepsis and lower in patients with baseline cUTI in the cefiderocol group compared with those in the BAT group. A higher mortality rate in the cefiderocol compared with the BAT group at day 49 was also observed among patients with A baumannii or P aeruginosa as a baseline pathogen and those with an acute physiology and chronic health evaluation (APACHE) II score ≥16. Although there were numerical differences in death at different time points between cefiderocol- and BAT-treated patients, confidence intervals for mortality differences did not provide nominally statistically significant evidence of a mortality increase at any time point.23
An independent adjudication committee determined that a greater percentage of patients in the cefiderocol group than in the BAT group had infection-related death with treatment failure (15.8% vs 8.2%) but also noted an imbalance in death due to underlying comor­bidities (9.9% vs 4.1%). Based on the findings from this trial, the following warning is included in its package insert: “Increase in All-Cause Mortality in Patients With Carbapenem-Resistant Gram-Negative Bacterial Infections: An increase in all-cause mortality was observed in Fetroja-treated patients compared to those treated with best avail­able therapy (BAT). Reserve Fetroja for use in patients who have limited or no alternative treatment options for the treatment of cUTI. Closely monitor the clinical response to therapy in patients with cUTI.”24
APEKS-NP25 (A baumannii, P aeruginosa, Escherichia coli, K pneumoniae, and S maltophilia in Nosocomial Pneumonia) was a phase 3 global, double-blind, randomized, active-controlled, noninferiority study of 300 adult patients with documented nosocomial pneumonia caused by GNB who received either cefiderocol (2 g every 8 hours over 3 hours) or meropenem (2 g every 8 hours over 3 hours). The primary end point was all-cause mortality at day 14 for the MITT population, defined as all treated patients except those that had gram-positive pathogens identified only at baseline. In the ITT population, 148 patients were random­ized to cefiderocol and 150 to high-dose meropenem.
With regard to the population at baseline, many patients (59.7%) were ventilated, prior antibiotic failure was commonplace (32.6%), and the median APACHE II score was 15. In contrast to the CREDIBLE-CR study, no difference in mortality was observed between cefid­erocol and meropenem in the ITT population at days 14 and 28, as well as at the end of study (12.8% vs 11.4%, 21.2% vs 20.1%, and 26.9% vs 22.8%, respectively). At day 14, all-cause mortality in the MITT population was 12.4% for cefiderocol and 11.6% for high-dose meropenem. In the microbiologically evaluable per protocol population, all-cause mortality was 12.4% for cefiderocol and 13% for high-dose meropenem. Mortality at day 14 was also comparable between cefiderocol and high-dose mero­penem among patients that had P aeruginosa at baseline (2 of 24 [8.3%] vs 3 of 24 [12.5%]), A baumannii at baseline (5 of 23 [21.7%] vs 4 of 24 [16.7%]), and an APACHE-II score ≥16 (12 of 71 [16.9%] vs 12 of 71 [16.9%]).
Clinical and microbiological outcomes were also compa­rable between groups at TOC. Notably, in the MITT popula­tion, 22.2% of patients in the meropenem arm were infected by GNB resistant to meropenem (MIC >8 μg/mL) compared with 1.7% of patients treated with cefiderocol who were infected by GNB potentially resistant to cefiderocol (MIC > 4 μg/mL). No unexpected safety signals were observed in the study, and the incidence of treatment-emergent adverse events was similar between treatment arms.
There is an urgent need for novel antimicrobial agents to address the emergence of MDR pathogens, an increasing cause of morbidity and mortality worldwide.26,27 Antibiotic development initiatives over the past several years have intro­duced new therapies with novel mechanisms that may show promise in the treatment of patients with infections due to highly resistant GNB. Based on the available data, IMI/REL and cefiderocol appear to be 2 promising treatments for these patients. Additional data are essential to further eluci­date the safety and efficacy of these agents in real-world settings, especially among patients with highly resistant gram-negative infections.
Lodise is a professor at Albany College of Pharmacy and Health Sciences in New York. He is also an infectious diseases clinical pharmacy specialist at Albany Stratton VA Medical Center. Bidell is a clinical pharmacist in infectious diseases and internal medicine at Massachusetts General Hospital in Boston. * indicates active member of the Society of Infectious Diseases Pharmacists.