The role of PK/PD–based strategies to preserve new molecules against multi-drug resistant gram-negative strains

In the last decade, infections caused by multi-drug resistant (MDR) Gram-negative (GN) strains have been increasing worldwide, especially in intensive care units (ICUs).¹ More than one third of hospital-acquired infections in ICUs are caused by carbapenem resistant Enterobacterales (CRE), Pseudomonas aeruginosa (PA) and Acinetobacter baumannii (AB) with a strong impact on hospital/ICU length of stay, mortality rate, and costs.^2,3 In the 2012-2017 period, infections caused by Extended spectrum β-lactamase (ESBL) producing strains increased by 53% in US hospitals.⁴ In the last few years, due to the lack of new/effective molecules, old ones such as colistin and fosfomycin, with a narrow therapeutic window and/or unpredictable PK, have come back into use.⁵

Recently, the availability of new drugs or new associations based on old or new β-lactam molecules with either a novel or a classic β-lactamase inhibitor (BL/BLI) have enabled clinicians to tackle MDR GN infections again. However, it is of primary importance to use these new therapeutic opportunities according to antimicrobial stewardship programs - which necessarily take into consideration pharmacokinetic (PK) and pharmacodynamic (PD) indexes - to limit the emergence of resistance against them.⁶

The emergence of resistance during antimicrobial therapy is particularly at risk in the following circumstances:

• when treating critically ill septic patients (due to the frequency of MDR strains and the changes in sepsis-related pharmacokinetics);

• when treating pulmonary infections (a difficult-to-reach infection site)

• when renal function is altered (this commonly occurs for these BL/BLI which are eliminated via the kidneys).

Here we will focus on the PK/PD data available for the latest 5 β-lactam drugs which target MDR GN strains, and we will provide some suggestions for better treatment and to prevent the emergence of resistance.

Ceftolozane/Tazobactam (TOL/TAZ)—the combination of a new fourth-generation cephalosporin with a former β-lactamase sulfone inhibitor has demonstrated strong activity against ESBL+ Enterobacterales and MDR PA.⁷ Both the molecules have a time-dependent activity, therefore the percentage time above MIC of the unbound drug (%fT > MIC) is the major PK/PD parameter that better describes their activity against MDR GN (Tables 1 and 2).^8–12 Among the new molecules/new associations, both ceftolozane and tazobactam are, together with meropenem vaborbactam (MER/VAB), those which best penetrate the epithelial lining fluid (ELF). After doubling the suggested standard dose and with 1 hour infusion, a higher concentration in ELF can be obtained and this can be effective against pneumonia caused by MDR PA. Recently, a non-inferiority phase 3 randomized controlled trial showed that high dosage TOL/TAZ is as efficient as meropenem in treating ventilation-associated pneumonia (VAP) caused by GN strains.¹³ Dose adjustment of TOL/TAZ is suggested in the presence of renal dysfunction.¹⁴ However, whether this reduction is advisable has been questioned after the results obtained in trials where there was a worse outcome in patients with reduced Creatinine Clearance (CrCl), for whom the BL dose had been decreased.¹⁵ TOL/TAZ extended infusion regimens have been proposed as a potential administration modality to improve the Probability Target Attainment (PTA) against infections due to MDR-PA with higher MIC values.¹⁶

Table 1. Pharmacokinetic (PK), Percentage of binding protein, PK/PD, Dosage, and market approval for Ceftolozane/Tazobactam (TOL/TAZ), Ceftazidime-Avibactam (CAZ/AVI), Meropenem-vaborbactam (MER/VAB) and Cefiderocol.

Drug	FDA/EMA Approval	% Protein Binding	Manufacturer’s Dosage	Betalactam activity*
TOL/TAZ	cIAIs cUTIs VAP	Low for both	1.5 g q 8 iv in 1h infusion 3 g q 8 (VAP) iv	> 3PBPs
CAZ/AVI	cIAIs cUTIs VAP	Low for both	2.5 g q8 iv in 2h infusion	< 2PBPs
MER/VAB	cIAI, cUTI, HAP, VAP, Bacteraemia	Low for Meropenem 33% Vaborbactam	4 g q8 iv in 3h infusion	> 3PBPs
IMI/REL	cUTIs (FDA)	20 %	1.25 g q 6 in 30 min infusion	> 3PBPs
CEFIDEROCOL	cUTIs (FDA)	40-60% (mostly albumin)	2 g q8 iv in 3h infusion	> 3PBPs

* Betalactams having a prevailing activity only on one penicillin binding protein (PBP) have low or almost absent bactericidal activity, suffer from a considerable inoculum effect with a massive bacterial biomass increase and a high endotoxin release. Carbapenems, as well as 4^th generation cephalosporins, due to their capability of inhibiting more than two main PBPs, generally show a minimal inoculum effect, no biomass increase and a negligible endotoxin release and therefore are associated always with a fast and high killing activity. Cephalosporins such as ceftazidime, being capable of inhibiting no more than 2 PBPs have intermediate results (8).

Legenda cIAIs = complicated intra-abdominal infections, cUTIs = complicated Urinary Tract Infections, VAP = ventilation associated pneumonia.

Table 2. Spectrum of activity and MIC breakpoints of Ceftolozane/Tazobactam (TOL/TAZ), Ceftazidime-Avibactam (CAZ/AVI), Meropenem-vaborbactam (MER/VAB) and Cefiderocol.and Imipenem/Relebactam (IMI/REL).

DRUGS	SENSITIVITY (S) MIC BREACKPOINTS (GENERIC)	PSEUDOMONAS AERUGINOSA XDR/PDR	ENTEROBACTERALES ESBL				ACINETOBACTER BAUMANNII
			E.COLI	KLEBSIELLA SPP	ENTEROBACTER	P. MIRABILIS
TOL/TAZ	S≤ 4 mg/mL	More active than CAZ/AVI	Active	Active	Active	Active	Low activity
			No activity on KPC, VIM, NDM, GES	No activity on KPC, VIM, NDM, GES	No activity on KPC, VIM, NDM, GES	No activity on KPC, VIM, NDM, GES
CAZ/AVI	S≤ 8 mg/mL	Less activity than TOL/TAZ	Active	Active	Active	Active	Low Activity
			No activity on Class B MBL	No activity on Class B MBL	No activity on Class B MBL	No activity on Class B MBL
MER/VAB	S≤ 8 mg/mL	Active	Active Activity on Class A/ C	Active Activity on Class A and C	Active Activity on Class A/C	Active Activity on Class A/ C	No further activity than meropenem alone
			No activity on Class B and D	No activity on Class B and D	No activity on Class B and D	No activity on Class B and D
IMI/REL	S ≤4 mg/L	Active	Active	Active	Active	Active	Not Active
			Activity on Class A and C β-lactamase	Activity on Class A and C β-lactamase	Activity on Class A and C β-lactamase	Activity on Class A and C β-lactamase
			No activity on Class B and OXA-48 class D β-lactamase	No activity on Class B and OXA-48 class D β-lactamase	No activity on Class B and OXA-48 class D β-lactamase	No activity on Class B and OXA-48 class D β-lactamase
Cefiderocol	S≤ 4 mg/mL	Active	Active Class A,B,C,D	Active Class A,B,C,D	Active Class A,B,C,D	Active Class A,B,C,D	Active

Legenda: ESBL Extended Spectrum Beta-Lactamase, KPC Klebsiella producing carbapenemase, MBL Metallo Beta- Lactamase, NDM New Delhi metallo-β-lactamase, VIM Verona integron-encoded metallo-β-lactamase, GES Guiana extended-spectrum beta-lactamases.

Ceftazidime-Avibactam (CAZ/AVI) is a combination of a third-generation cephalosporin and avibactam, a novel diazabicyclo-octane β-lactamase inhibitor.¹⁷ The presence of avibactam enables ceftazidime to recover its activity versus ESBL and Class A, C and some D (OXA 48), but not Class B (metallo-β-lactamase, MBL) carbapenemases¹⁸ (Table 3). For ceftazidime, the more important PK/PD index is represented by fT > MIC, while for avibactam both fT > MIC and AUC/MIC are important.¹⁹ Indeed, for Enterobacterales and PA, the optimal PK/PD target is 50% fT > MIC for ceftazidime and a concentration threshold >1 mg/L (with consequent AUC₀-₂₄, 48 mg/l·h) for avibactam.^18,20 However, avibactam concentrations >4 mg/L are necessary in order to reduce the bacterial load of KPC-producing Klebsiella pneumoniae (KP), due to low efficacy of this BLI on KPC 3 and 4.^20,21 Ceftazidime and avibactam enter the ELF to the same degree, with average lung AUC values being 30% of those obtained in plasma.²² While these concentrations have been shown to generally exceed those effective in experimental pneumonia models, to date there is no evidence that similar exposures can be obtained among critically-ill patients, in whom concentration in the ELF can be decreased during pneumonia for several reasons, including changes in pulmonary circulation.²³ Consistently, Shields et al., when they analyzed the risk factors for CAZ-AVI failure in patients with CRE infections, found positive associations with healthcare-associated pneumonia (HAP) and renal replacement therapy (RRT) (OR 3.09, 95% CI 1.03-9.34 and 4.78, 95% CI 1.03–22.2, respectively).²⁴ The latter was also a risk factor for the emergence of resistance (OR 26.67, 95% CI 26.67– 317.1) (OR 3.09, 95% CI 1.03–9.34). It cannot be excluded that resistance development and treatment failure in the population analyzed by Shields were due to inappropriate dosing. It has been proposed that patients receiving CAZ/AVI should be ‘aggressively dosed’ in order to obtain optimal exposures at least for the first 48 hours of treatment, irrespective of their renal function.²³

Table 3. Pharmacokinetic (PK)/Pharmocodynamic (PD) PK/PD, Dosage adjustment, Infusion needs for Ceftolozane/Tazobactam (TOL/TAZ), Ceftazidime-Avibactam (CAZ/AVI), Meropenem-vaborbactam (MER/VAB) and Cefiderocol.

Drug	PK/PD target	Prolonged Infusion	Dose Adjustment according to renal function	Suggestion for conservation in severe infections /difficult site of infections
TOL/TAZ	32% fT > MIC	Suggested	Yes if CrCl < 50	Measure CrCl daily with Chen formula Evaluate if dosing reduction is really needed in the first 48 h necessary TDM
CAZ/AVI	50% fT > MIC	Suggested	Yes if CrCl < 50	Consider infection site: if pulmonary: increased dose TDM if possibile Combination therapy con amikacin or colistin^9^,^10 Combination therapy with aztreonam^11^,^12
MER/VAB	40% fT > MIC	Suggested	Yes if CrCl < 50	Vaborbactam has a weak activity on OXA-48-like TDM
IMI/REL	69% fT > MIC (I) fAUC0–24 hours/MIC ≥ 51.9 (R)	Suggested	Yes if CrCl < 50	TDM
Cefiderocol	75% fT > MIC	Suggested	Yes if CrCl < 50	TDM

Legenda: MIC minimum inhibitory concentration, CrCl creatinine clearence, TDM Target Drug Monitoring.

Meropenem-vaborbactam (MER/VAB) is a combination of a broad spectrum carbapenem and vaborbactam, a boronic acid β-lactamase inhibitor with no intrinsic antibacterial activity. Vaborbactam is unique in that it is the first cyclic boronic acid inhibitor with high affinity for serine residues, allowing it to serve as a competitive inhibitor, by forming a covalent bond with beta-lactamases, without undergoing hydrolysis.²⁵ Vaborbactam enables meropenem to be effective against Enterobacterales which produce KPC and other class A serine carbapenemases, but not against class D or class B carbapenemases.²⁶ MER/VAB potency is highly reduced in KPC-producing KP overexpressing the major multidrug efflux pump AcrAB-TolC and lacking both ompK35 and ompK36 porins, used by MER/VAB to cross the outer membrane of KP.²⁷ Moreover, this combination does not have any further activity than the meropenem alone towards PA, since PA resistance is principally related to outer membrane impermeability, upregulation of efflux systems, and production of class B or class D β-lactamases.⁵ Meropenem is a time-dependent antimicrobial agent and benefits from prolonged infusions in order to keep a concentration > MIC for 40% of the dosing interval (fT > MIC> 40%), a value which is necessary to kill bacteria.^28,29 In order to be effective, vaborbactam needs to maintain a concentration ≥8 mg/L in plasma and its PK/PD efficacy is connected to the ratio of vaborbactam AUC and the MIC, due to its rapid target binding and its slow reversibility.³⁰ In vitro data seem to indicate that selection of resistance is less common with MER/VAB than with other newBL/BLI, but more robust clinical data are needed to confirm this. Moreover, caution should be used in the treatment of susceptible CRE when OXA-based enzymes have not been excluded.²³

Imipenem/Relebactam (IMI/REL) is a combination of the carbapenem imipenem and relebactam, a diazabicyclooctane beta-lactamase inhibitor with a similar structure of avibactam but with an additional piperidine ring.³¹ Relebactam has been shown to be efficient against serine β-lactamase class A and C and some class D. It has no effect on class B and OXA48 class D. Imipenem (always administered with cilastatin) is a time-dependent antibiotic and a prolonged infusion is necessary in order to obtain at least a target of fT > MIC 69% for bacteriostasis.³¹ In order to be effective, relebactam needs to have a concentration of 4 mg/L and the PK/PD index predicting efficacy is fAUC/MIC, with a target value of 7.5.³² IMI/REL has recently been approved for the treatment of cUTIs and IAIs. While IMI/REL shows potential restoration of activity in isolates resistant to imipenem/cilastatin, much less data are available on its role in clinical practice, including its stability against the development of resistance. In relebactam, desulfation (and consequent inactivation) occurs more slowly than in avibactam, but relebactam shows a reduced inhibitory potency than avibactam (about 1/8).³³ This combination, unlike the other BLI/BLI, is administered as a 30 min infusion. The optimal duration of administration has not yet been validated in the clinical setting.

Cefiderocol is a fourth-generation cephalosporin containing a Cl-cathecol side chain with a chemical structure related to ceftazidime and cefepime but with less susceptibility to beta-lactamases such as AMP C, KPC, NDM and OXA carbapenemases and ESBLs. Cefiderocol is therefore active against KPC, and NDM carbapenemases, MDR PA, AB and Stenotrophomonas maltophilia. This molecule creates a complex with ferric iron and promotes a siderophore-like transport across the outer membrane of GN, behaving like a ‘Trojan Horse’.³⁴ Once in the periplasmic space, its separation from ferric iron enables the molecule to impair peptidoglycan synthesis by binding to penicillin-binding proteins (PBPs).^35–37 Cefiderocol elimination is directly related to renal function.^38,39 A 90% PTA can be obtained with a dose of 2 g x 3 in 3 hours of infusion, with a target of 75% fT > MIC.⁴⁰ This antimicrobial has been approved for cUTIs, but it could be effective in HAP, where MDR Gram-negative bacilli are responsible, since an adequate concentration in the ELF has been shown in healthy volunteers. Clinical trials testing the possible use of the molecule for pneumonia are ongoing.^23,41

Reduced in vitro activity of cefiderocol has been seen towards PA strains that show a decrease in the production of iron transport system components or mutations on the outer membrane for the iron transport system.^42,43 At present it is not known whether in patients with low free iron concentrations the development of resistance to this molecule is likely. A 90% PTA for bloodstream and urinary GN bacteria with cefiderocol MICs ≤ 4 mg/L is obtainable for the dose of 2 g IV every 8 hours as a 3-hr infusion in patients with normal renal function.⁴⁰ Suggestions for a 6-hr interval for patients with increased renal function has been proposed. Cefiderocol may be the first monotherapy for treating metallo beta-lactamases.

Suggestions for conserving new beta-lactams in severely ill patients

New or recently introduced beta-lactam molecules are available to treat infections caused by MDR GN bacteria. Since a rapid emergence of resistance has been described/reported for some of them, the judicious use of these drugs with the aim of preserving them from resistance is a high priority.

For some of these beta-lactam molecules their true specific clinical role together with the possible contribution of an associated therapy have not yet been defined. Moreover, ideal dosing strategies, especially for fragile/severely ill patients (such as ICU patients), are still lacking. Rapid bedside diagnostic tests to identify pathogens and detect their mechanisms of resistance will hopefully be available in the near future, together with real-world clinical data, which will help to provide evidence concerning their efficacy and risk of developing resistance.

Three Tables are reported below that compare these 5 new antibiotics according to their PK characteristics, spectrum of activity, PK/PD targets, susceptibility breakpoints and clinical use. Some suggestions for their use according to PK/PD parameters, with the aim of optimizing clinical outcome and containing the risk of resistance, are also reported.

All 5 molecules belong to the beta-lactam antibiotic class and share some characteristics. Being hydrophilic molecules, there is the risk that if prescribed according to the manufacturer’s dose they are under-dosed in critically ill septic patients, in whom permeability and volume of distribution are increased.⁴⁴ The risk of underexposure (which affects clinical cure and favors resistances) must be considered in ICU patients in whom renal function is altered. Renal function can be not only reduced, but also increased in a certain subset of ICU patients; moreover, kidney function is not static, and changes (both increasing and decreasing) can occur very rapidly.¹⁵ Therefore, increased doses during increased renal clearances should be proposed, while even in the case of reduced renal function some authors suggest not reducing doses, at least for the first 48 hrs of treatment.¹⁵ A keen revision of dosing according to CrCl formulas validated for ICU patients must also be applied. Finally, data on the probability of reaching PK/PD targets in ICU patients undergoing RRT are also necessary.

In order to limit the development of resistance there is the need for specific PK/PD studies, on specific subsets of patients, that can validate both the use of these molecules in deep-seated infections and, for some of them, the value of association therapy to limit the possibility of emergence of resistance. Finally, with the aim of customizing/individualizing antimicrobial therapy, the use of bedside therapeutic drug monitoring (TDM), possibly in a simplified, clinically affordable modality,⁴⁵ should be encouraged as part of routine care in the near future.

Collectively, these efforts may help to preserve these new drugs for a longer period.

Conflicts of interest

The authors report no conflict of interest.

Additional information

Notes on contributors

Chiara Adembri

Doctor Iacopo Cappellini MD, PhD, Anaesthesiologist, expert in Critical Care Medicine and infections in the critically ill patient.

Iacopo Cappellini

Prof Chiara Adembri MD, PhD, Anaesthesiologist, expert in antimicrobial therapy and infections in the critically ill patient.

Andrea Novelli

Prof Andrea Novelli, MD, pharmacologist, expert in antimicrobial therapy and PK/PD applied to antimicrobial therapy.