Pharmacokinetics/Pharmacodynamics of Vaborbactam, a Novel Beta-Lactamase Inhibitor, in Combination with Meropenem

Vaborbactam is a novel beta-lactamase inhibitor with activity against important beta-lactamases, in particular, serine carbapenemases, and is currently approved in combination with meropenem as Vabomere for the treatment of complicated urinary tract infections, including pyelonephritis. This combination is highly active against Gram-negative pathogens, especially Klebsiella pneumoniae carbapenemase (KPC)-producing carbapenem-resistant Enterobacteriaceae.

ABSTRACT

Vaborbactam is a novel beta-lactamase inhibitor with activity against important beta-lactamases, in particular, serine carbapenemases, and is currently approved in combination with meropenem as Vabomere for the treatment of complicated urinary tract infections, including pyelonephritis. This combination is highly active against Gram-negative pathogens, especially Klebsiella pneumoniae carbapenemase (KPC)-producing carbapenem-resistant Enterobacteriaceae. The objective of these studies was to evaluate vaborbactam pharmacokinetics (PK) and pharmacodynamics (PD) relationships for efficacy in a neutropenic mouse thigh infection model, as well as in an in vitro hollow-fiber infection model, in combination with a fixed exposure of meropenem using KPC-containing strains of Enterobacteriaceae. For both models, the meropenem dosage regimen was designed to simulate a 2-g dose administered every eight hours (q8h) by 3-h infusion. Vaborbactam dosage regimens were designed to produce a wide range of 24-h areas under the concentration-time curves (AUCs) in the thigh infection model. However, for the hollow-fiber model, the AUCs were limited to values of 192, 320, or 550 mg · h/liter. In both the animal and in vitro models, the PK-PD parameter that best described the antibacterial activity of vaborbactam, when administered in combination with meropenem at exposures equivalent to 2 g dosed q8h by 3-h infusion in humans, was the 24-h free vaborbactam AUC/meropenem-vaborbactam (with vaborbactam at 8 mg/liter) MIC ratio. The magnitude of this ratio for bacteriostasis was 9 to 12 and the magnitude to observe a 1-log kill was 18 to 38. In addition, a magnitude greater than 24 suppressed the development of resistance in the in vitro hollow-fiber model.

INTRODUCTION

The pharmacokinetics (PK) and pharmacodynamics (PD) of numerous antibacterial agents have been studied using animal and in vitro models of infection. These models enable the assessment of the exposure-response relationships required to produce an antibacterial effect and the magnitude of that effect against their respective target organisms. The PK-PD relationship that best describes the antibacterial effect and the magnitude of that effect has been described most often using the neutropenic mouse thigh infection model (1). The magnitude of the relationship determined in this model has been found to be similar to that necessary for efficacy in humans (2).

Meropenem is a broad-spectrum intravenous carbapenem antibiotic used to treat a wide variety of infections. The spectrum of action includes many Gram-positive, Gram-negative, and anaerobic bacteria. Meropenem is approved in many countries around the world at doses up to 2 g every 8 h (q8h).

The PK-PD relationship for meropenem has been studied extensively in vitro, in animals, and in humans (3–9). Since meropenem has time-dependent bactericidal activity, the relationship that best describes the antibacterial activity of meropenem is the percentage of the dosing interval that free drug levels exceed the MIC, or %T_>ƒMIC. The magnitude that best correlates with efficacy in vitro, in vivo, and in humans is a %T_>ƒMIC of 40 (10). To maximize the exposure of meropenem, as part of meropenem-vaborbactam combination (Vabomere), a dose of 2 g administered q8h by 3-h infusion was selected. With this dosage regimen, free meropenem plasma levels are expected to exceed 8 mg/liter for at least 40% of the dosage interval on the basis of a 10,000-patient Monte Carlo simulation (11–13).

Vaborbactam is a novel beta-lactamase inhibitor that was developed to inhibit the Klebsiella pneumoniae carbapenemase (KPC) beta-lactamase (14). Thus, vaborbactam restores the activity of meropenem against KPC-producing carbapenem-resistant Enterobacteriaceae (CRE), a pressing public health threat (15).

As vaborbactam is a novel beta-lactamase inhibitor, the exposure-response relationship associated with maximum potentiation and resistance suppression, in combination with meropenem for Enterobacteriaceae strains expressing KPC, has not been determined. The objective of these studies was to evaluate vaborbactam PK-PD relationships for efficacy, in combination with a fixed exposure of meropenem equivalent to 2 g q8h by 3-h infusion in humans using KPC-containing strains of Enterobacteriaceae, in both the neutropenic mouse thigh infection and in vitro hollow-fiber models.

(This work was presented in part at the 52nd and 54th Interscience Conferences on Antimicrob Agents Chemother, 2012 and 2014, respectively.)

RESULTS

MIC testing.

The pretreatment MIC values for the strains used in these studies are shown in Table 1. All strains produced KPC-2 or KPC-3 and had meropenem MICs of 8 mg/liter or greater. In addition, some strains had mutations in outer membrane porins OmpK35 and OmpK36, which have an impact on the uptake of carbapenems (16).

TABLE 1.

MICs and characteristics of Enterobacteriaceae strains used in mouse and in vitro PK-PD studies

Strain	Beta-lactamase(s)	OmpK35a	OmpK36a	Meropenem MIC (mg/liter)		Model
				Alone	With vaborbactam 8 mg/liter
E. coli EC1007	KPC-3	ND	ND	8	≤0.06	HFb
E. cloacae ECL1058	KPC-3, SHV-11, TEM-1	FL	FL	8	0.125	HF
E. cloacae ECL1061	KPC-3 (hyper AmpC expression)	FS aa no. 287	FL	16	0.125	HF
E. cloacae ECL1079	KPC-3	Stop aa no. 60	Stop aa no. 77	>64	8	Mouse, HF
K. pneumoniae KP1061	KPC-3, SHV-11, TEM-1	FS aa no. 42	FL	16	≤0.06	HF
K. pneumoniae KP1087	KPC-2, CTX-M-15, SHV-11, TEM-1	FS aa no. 208	GD	32	0.25	HF
K. pneumoniae KP1074	KPC-3, SHV-11, TEM-1	FS aa no. 42	GD	>64	0.5	HF
K. pneumoniae KP1093	KPC-3, SHV-11, TEM	FS aa no. 42	GD	>64	0.5	Mouse, HF
K. pneumoniae KP1099	KPC-2, SHV-11, SHV-12, CTX-M-14	FS aa no. 29	GD	>64	1	HF
K. pneumoniae KP1094	KPC-2, TEM-1, LEN-17	Stop aa no. 230	Stop aa no. 92	>64	4	Mouse, HF
K. pneumoniae KP1100	KPC-3, TEM, SHV	FS aa no. 42	GD	>64	4	HF
K. pneumoniae KP1194	KPC-2 TEM SHV	FS aa no. 42	IS at −45	>64	8	HF
K. pneumoniae KP1223	KPC-2, SHV, TEM	FS aa no. 29	GD	>64	8	Mouse, HF
K. pneumoniae KP1096	KPC-2, TEM, SHV-11	L63V, E132K	IS at nt no. 126	>64	16	Mouse, HF
K. pneumoniae KP1244	KPC-3, SHV-11, SHV-12	FS aa no. 42	R191L, T333N	>64	16	HF
K. pneumoniae KP1092	KPC-2, SHV-11, SHV-12, TEM-1	FS aa no. 42	IS at −45	>64	32	HF
K. pneumoniae KP1254	KPC-2, SHV, TEM, OXA-10	FS aa no. 42	IS and ΔopmK36	>64	64	HF

^aND, not determined; FL, full length (functional); stop aa no., nonsense mutations resulting in a truncated nonfunctional protein; FS aa no., frameshift mutation resulting in a nonfunctional protein; GD, insertion of two amino acids (Gly134Asp135) resulting in a narrow semifunctional channel; IS at −45, promoter insertion of insertion sequence element resulting in downregulation of expression; nt, nucleotide; T333N and E132K are loss-of-function mutations in ompK36 of KP1096 and KP1244, respectively.

^bHF, hollow fiber.

Mouse pharmacokinetic parameters.

The plasma pharmacokinetic parameters for meropenem alone, vaborbactam alone, or both drugs in combination in neutropenic mice have been described previously (17). Briefly, a dose of 300 mg/kg meropenem given every 2 h for 24 h produced a similar free drug time at an MIC of >8 mg/liter as a dose of 2 g administered q8h by 3-h infusion in humans. A dose of 50 mg/kg vaborbactam given every 2 h for 24 h produced a similar free drug 24 h vaborbactam area under the concentration-time curve (AUC) as the dose of 2 g administered q8h by 3-h infusion in humans (Table 2).

TABLE 2.

Comparison of the pharmacokinetics of meropenem and vaborbactam in mice and in humans

Compound	Species	Dosage regimen	24-h free AUC (mg · h/liter)	%T>8 mg/liter
Meropenem	Human	2 g q8h by 3-h infusion	402	56
	Mouse	300 mg/kg q2h	1,572	51
	Human	1.5 g q8h by 3-h infusion	282	47
	Mouse	200 mg/kg q2h	1,080	47
	Human	1 g q8h by 3-h infusion	162	38
	Mouse	100 mg/kg q2h	588	39
Vaborbactam	Human	4 g q8h by 3-h infusion	686	100
	Mouse	100 mg/kg q2h	720	70
	Human	2 g q8h by 3-h infusion	343	72
	Mouse	50 mg/kg q2h	360	53
	Human	1 g q8h by 3-h infusion	172	44
	Mouse	25 mg/kg q2h	180	30
	Human	500 mg q8h by 3-h infusion	86	24
	Mouse	12.5 mg/kg q2h	90	18
	Human	250 mg q8h by 3-h infusion	43	0
	Mouse	6.25 mg/kg q2h	45	0

Determination of vaborbactam PK-PD using the neutropenic mouse thigh infection model.

The pharmacodynamics (PD) of vaborbactam were studied in combination with a fixed exposure of meropenem in a neutropenic mouse thigh infection model against carbapenem-resistant K. pneumoniae and Enterobacter cloacae strains. In these studies, untreated control organisms grew by 1 to 3 logs compared to that of the untreated controls at the start of treatment. For all strains, a regimen of meropenem alone, similar to 2 g q8h by 3-h infusion in humans, had either bacteriostatic effects or a lack of bacterial killing. However, the combination of meropenem plus vaborbactam produced bacterial killing against all strains with meropenem-vaborbactam MICs ranging from 1 to 16 mg/liter.

The data for the four K. pneumoniae isolates and the single E. cloacae isolate tested in the neutropenic mouse thigh infection model were pooled for PD analysis. The relationships between each measured parameter and changes in log CFU/thigh are shown in Fig. 1. The magnitudes of the relationship required to achieve stasis, 1 log of bacterial killing (1-log kill), and 2 logs of bacterial killing (2-log kill) are provided in Table 3. Among the PK-PD indices evaluated in the thigh infection model, the ratio of the 24-h free vaborbactam AUC/meropenem-vaborbactam MIC provided the best overall fit for the data. The magnitude of the free 24-h AUC/meropenem-vaborbactam MIC required for a bacteriostatic effect or for 1 log of bacterial killing was 9 or 38, respectively.

FIG 1.

Vaborbactam PK-PD in the neutropenic mouse thigh infection model. VAB, vaborbactam; M-V, meropenem-vaborbactam.

TABLE 3.

PK-PD parameters and corresponding magnitudes required for effect in the neutropenic mouse thigh infection model

PK-PD parameter	Magnitude required for:
	Stasis	1-log kill	2-log kill
%Free > 4 mg/liter	21	54	95
%Free > 8 mg/liter	12	35	62
Free 24-h AUC	50	267	720
Free 24-h AUC/meropenem-vaborbactam MIC	9	38	220

In vitro hollow-fiber pharmacokinetic parameters.

All the targeted PK values were simulated on the basis of clinical studies and have been described previously (16). Briefly, meropenem levels were >8 mg/liter for 75% of the dosing interval and >16 mg/liter for 40% to 50% of the dosing interval, and the vaborbactam 24-h AUCs ranged from 192 to 550 mg · h/liter.

Determination of vaborbactam PK-PD using an in vitro hollow-fiber pharmacodynamic model.

The PD of vaborbactam were also studied in combination with a fixed exposure of meropenem in an in vitro hollow-fiber pharmacodynamic model using humanized dosage regimens against a total of 17 KPC-producing isolates: thirteen clinical isolates of K. pneumoniae, three clinical isolates of E. cloacae, and one clinical isolate of Escherichia coli.

In contrast to the mouse model data, the starting inoculum used in all of the in vitro hollow-fiber PK-PD studies was ∼10⁸ CFU/ml. By using a high inoculum, the objectives were to not only determine the linked PK-PD parameter but also determine the magnitude of that parameter required to suppress resistance development and prevent regrowth.

The data from multiple hollow-fiber studies using 17 strains were pooled for PD analysis. The only relationship that could be fit to an Emax model was the change in log CFU/ml and the 24-h free vaborbactam AUC/meropenem-vaborbactam MIC ratio (Fig. 2). The magnitudes of the relationship required to achieve stasis, 1 log of bacterial killing, 2 logs of bacterial killing, 3 logs of bacterial killing, and suppression of resistance development are shown in Table 4. The magnitude of the PK-PD targets (24-h free AUC/MIC) required for a bacteriostatic effect or for 1 log of bacterial killing were 12 or 18, respectively.

FIG 2.

Vaborbactam PK-PD in the in vitro hollow-fiber infection model. VAB, vaborbactam; M-V, meropenem-vaborbactam.

TABLE 4.

PK-PD parameters and magnitude required for effect in the hollow fiber model

PK-PD parameter	Magnitude required for:
	Stasis	1-log kill	2-log kill	3-log kill	Resistance prevention
%Free > 4 mg/liter	No relationship found
%Free > 8 mg/liter
Free 24-h AUC
Free 24-h AUC/meropenem-vaborbactam MIC	12	18	25	36	>24

To determine the vaborbactam exposure required to suppress resistance development in the background of 2 g meropenem q8h by 3-h infusion, K. pneumoniae strains KP1096, KP1194, and KP1223, with meropenem-vaborbactam MICs of 16, 8, and 8 mg/liter, respectively, were studied using 24-h vaborbactam AUCs of 192, 320, and 550 mg · h/liter. As shown in Fig. 3, these strains produced resistant mutants when the ratio of the 24-h free vaborbactam AUC to the meropenem-vaborbactam MIC was less than 24. The resistant mutants had a 4-fold increase in the meropenem-vaborbactam MIC compared to pretreatment values. MICs increased from 8 to 16 mg/liter to 64 mg/liter.

FIG 3.

Relationship between change in log CFU/ml, free 24-h vaborbactam AUC/M-V MIC, and resistance suppression. M-V, meropenem-vaborbactam.

DISCUSSION

The PK-PD of numerous antimicrobial agents have been studied using animal and in vitro models of infection. These models enable the assessment of the exposure-response relationships required to produce an antibacterial effect and the magnitude of that effect against their respective target organisms. The study of the PK-PD of different beta-lactamase inhibitors in combination with beta-lactams has increased recently with different pharmacodynamic parameters proposed for each one. The PD parameter that best correlated with efficacy for avibactam in the presence of ceftazidime was found to be the percentage of the dosing interval that free drug levels exceeded 1 mg/liter (18). For tazobactam, with either ceftolozane or piperacillin, the PD parameter that best correlated with efficacy was found to be the percentage of the dosing interval that free drug levels exceeded 50% of the potentiated MIC (with tazobactam at 4 mg/liter) (19, 20). Another beta-lactamase inhibitor, CB-618, was studied in combination with cefepime, ceftazidime, ceftolozane, and meropenem. For CB-618, the PD parameter that best correlated with efficacy was found to be the 24-h free AUC/potentiated MIC ratio (21). The goals of our studies were to determine the PD parameter for vaborbactam that best correlated with meropenem-vaborbactam efficacy and to determine the magnitude of that parameter that produced bacterial killing and suppressed the development of resistance against KPC-producing carbapenem-resistant Enterobacteriaceae strains.

To determine and quantify the pharmacodynamic parameters of vaborbactam, a fixed exposure of meropenem was combined with various exposures of vaborbactam in both animal and in vitro models of infection.

Both the neutropenic mouse thigh infection and the in vitro hollow-fiber infection models showed that the PK-PD parameter that best described the antibacterial activity of vaborbactam, when administered in combination with meropenem exposures equivalent to 2 g meropenem q8h by 3-h infusion in humans, was the 24-h free vaborbactam AUC/meropenem-vaborbactam MIC ratio.

In these models, the magnitude of the 24-h free vaborbactam AUC/meropenem-vaborbactam MIC ratio required for a bacteriostatic effect was 9 and 12 in the neutropenic mouse and hollow-fiber models, respectively. The magnitudes of that same parameter required to produce 1 log of bacterial killing were 38 and 18 in the neutropenic mouse and hollow-fiber models, respectively.

Only a limited number of studies have focused on finding the optimal exposures required to suppress the amplification of less susceptible bacterial subpopulations (22). Traditionally, pharmacodynamic studies use a starting inoculum of 10⁵ to 10⁶ CFU/ml. In our hollow-fiber studies, we used an inoculum of 10⁸ CFU/ml to determine the optimal exposure necessary to suppress and/or minimize the development of resistance. Our studies found that a 24-h free vaborbactam AUC/meropenem-vaborbactam MIC ratio of 24 or greater is required to suppress the development of resistance.

Overall, these data show that a 2-g dose of vaborbactam administered every 8 h by 3-h infusion, in a fixed background of meropenem, produces an exposure that is sufficient to produce bacterial killing and suppress the resistance of KPC-producing carbapenem-resistant strains of Enterobacteriaceae with a meropenem-vaborbactam MIC of up to 8 mg/liter. These data support the FDA-approved dosage regimen of Vabomere, which is a vaborbactam dose of 2 g administered in combination with meropenem 2 g every 8 h by 3-h infusion.

MATERIALS AND METHODS

Antimicrobial agents.

Meropenem (lot no. DF-3297) was obtained from commercial sources, reconstituted as described in the prescribing information, and then diluted in saline to achieve desired concentrations. Vaborbactam (lot no. P-232-159-2) was synthesized at The Medicines Company and was reconstituted and further diluted in saline to achieve the required concentrations. For animal and in vitro combination therapies, meropenem and vaborbactam were admixed in saline and Mueller-Hinton broth (MHB), respectively.

Bacterial strains and MIC testing.

Four clinical isolates of K. pneumoniae and one clinical isolate of E. cloacae were used in a neutropenic thigh infection model. Thirteen K. pneumoniae isolates, three clinical isolates of E. cloacae, and one E. coli isolate were used for the in vitro hollow-fiber model study. Strains were grown in MHB at 37°C under constant aeration for 20 h. The inoculum used for infection was prepared by removing an aliquot that was subcultured into fresh MHB and allowed to regrow at 37°C under constant aeration. The suspension was incubated for 3 h to reach an absorbance of 0.30 to 0.35 at 600 nm. The bacterial suspensions were diluted in fresh MHB or saline to yield ∼10⁷ or 10⁸ CFU/ml by correlation of absorbance at 600 nm with predetermined plate counts. The pre- and posttreatment MIC values were determined for meropenem and meropenem-vaborbactam using the broth microdilution method as recommended by the Clinical and Laboratory Standards Institute (CLSI) (23).

Pharmacokinetics.

Pharmacokinetic studies in mice were conducted and analyzed as described previously (17). Briefly, neutropenic mice were administered meropenem and vaborbactam alone or in combination by the intraperitoneal route. At designated time points, the mice were euthanized and their blood was collected by cardiac puncture and transferred to EDTA-containing tubes. Blood samples were centrifuged within 5 min of collection at 12,000 × g for 5 min to obtain plasma. An equal volume of 3-(N-morpholino)propanesulfonic acid (MOPS) buffer (pH 7) was added to plasma samples that contained meropenem, which were then stored at −80°C until analyzed.

Determination vaborbactam PK-PD using a neutropenic mouse thigh infection model.

All studies using animals were performed under protocols approved by an Institutional Animal Care and Use Committee. Female Swiss Webster mice were rendered neutropenic and infected as described previously (17). Briefly, the mice were infected by intramuscular injections of 0.1 ml of inoculum (10⁷ CFU/ml) into both thigh muscles while under isoflurane anesthesia (5% isoflurane in oxygen running at 4 liter/min). Treatment regimens were chosen to simulate exposures found in humans. Meropenem administered at 300 mg/kg every 2 h over a 24-h period in mice produces an exposure equivalent to 2 g of meropenem administered every 8 h by 3-h infusion in humans. Vaborbactam administered at unit doses of 6.25 mg/kg, 12.5 mg/kg, 25 mg/kg, 50 mg/kg, or 100 mg/kg produce exposures equivalent to human doses of 0.25 g, 0.5 g, 1 g, 2 g, or 4 g, respectively, of vaborbactam administered every 8 h by 3-h infusion in humans (17). Doses were administered every 2 h by the intraperitoneal route. For each strain, untreated mice were euthanized prior to the start of treatment in order to determine baseline bacterial counts. All treatment groups were euthanized by carbon dioxide asphyxiation 2 h after the last dose administered. The thighs were removed aseptically and homogenized (Pro200 homogenizer; Pro Scientific, Monroe, CT) in sterile saline. The homogenized thighs were then serially diluted 10-fold and were plated on Mueller-Hinton agar (MHA). Colonies were counted the next day after the plates were allowed to incubate overnight at 37°C.

Determination vaborbactam PK-PD using an in vitro hollow-fiber model.

The in vitro pharmacodynamic model consists of central and peripheral compartments (24). The peripheral compartment consists of artificial capillary units (Cartridge C-3001; FiberCell Systems, Inc., Frederick, MD) arranged in series with the central compartment (25). Each capillary unit has a bundle of small semipermeable fibers with a molecular size retention of ca. 5,000 molecular weight (MW) to enable the passage of nutrients but restrict the passage of bacteria. The entire system was set up in a dry heat incubator adjusted to 37°C and included six artificial capillary units so that three strains could be tested in duplicates. Both the central and peripheral compartments were filled with cation-adjusted MHB. Each peripheral compartment (capillary unit and tubing) contained ca. 23 ml of growth medium.

Prior to the initiation of treatment, 1.5 ml of log-phase bacterial suspensions (∼10⁸ CFU/ml) were introduced into the peripheral chamber of the model and allowed to grow for 2 h. Doses were administered into the central compartment by 3-h infusions and circulated to peripheral hollow-fiber units by peristaltic pumps (MasterFlex L/S; Cole-Parmer Instrument Co., Vernon Hills, IL).

The dosage regimens studied were designed to simulate the human pharmacokinetic profiles of meropenem and vaborbactam administered every 8 h by 3-h infusion (26). For all studies, the dosage regimen for meropenem was fixed to simulate a 2-g dose administered every 8 h by 3-h infusion or a 24-h AUC of 420 mg · h/liter and provide concentrations >8 mg/liter for 50% of the dosage interval. The dosage regimens for vaborbactam followed the same half-life and clearance, but the AUC was varied from 192 mg · h/liter to 550 mg · h/liter.

For CFU counts, samples were collected from each artificial capillary unit at designated time points, serially diluted, and plated on MHA plates. The plates were incubated overnight at 37°C. The limit of detection was 100 CFU/ml. Meropenem and vaborbactam concentrations were determined by collecting samples from the central compartment at various time points over a 32-h period. An equal volume of MOPS buffer (pH 7) was added to samples to stabilize meropenem. The samples were stored at −80°C until analyzed.

Bioanalytical assays.

The bioanalytical assays for meropenem and vaborbactam have been described previously (16, 17). Briefly, standard curves were prepared in plasma or MHB. Twenty-five-microliter aliquots of sample were placed in 1.5-ml microcentrifuge tubes containing 200 µl of 4.0 µg/ml doripenem (internal standard for meropenem) and 4.0 µg/ml of RPX7015 (internal standard for vaborbactam) in 10%, 45%, 45% water-methanol-acetonitrile (vol/vol/vol). The samples were mixed using a vortex mixer and then centrifuged for 10 min at 15,000 × g using a tabletop centrifuge. The supernatant (∼150 µl) was removed and added to 400 µl of water in a 96-well plate. The samples were further mixed using a vortex mixer. Twenty microliters of each sample was injected onto a high-pressure liquid chromatography-mass spectrometer (HPLC-MS) for quantification. The lower limit of quantitation for both meropenem and vaborbactam was 0.04 µg/ml.

Pharmacokinetic and pharmacodynamic modeling.

Plasma (mouse model) or MHB (in vitro model) concentrations were fit using a one-compartment first-order model (Phoenix WinNonlin version 6.4; Certara USA, Inc., Princeton, NJ).

For the neutropenic mouse thigh infection model, the data for the four K. pneumoniae isolates and the E. cloacae isolate were pooled for the analysis. These pooled data were used to determine the relationship between the change in log CFU/thigh and the percentage of 24-h free vaborbactam concentrations that exceeded 4 mg/liter (%ƒvaborbactam concentration >4 mg/liter), %ƒvaborbactam concentration >8 mg/liter, 24-h free vaborbactam AUC, and 24-h free vaborbactam AUC/meropenem-vaborbactam (M-V) MIC ratio. The data for the thirteen K. pneumoniae isolates, the three E. cloacae isolates, and one E. coli isolate tested in the in vitro hollow-fiber PK-PD model were pooled for the analysis. As with the animal model data, these pooled data were used to determine the relationship between the change in log CFU/thigh and %ƒvaborbactam concentration >4 mg/liter, %ƒvaborbactam concentration >8 mg/liter, 24-h free vaborbactam AUC, and 24-h free vaborbactam AUC/M-V MIC.

The relationships between PK-PD parameters were analyzed using a sigmoid Emax pharmacodynamic model (Phoenix WinNonlin version 6.4; Certara USA, Inc., Princeton, NJ) using the following equation: reduction in log CFU/thigh or ml = [(E_max × Xγ)/(EC₅₀γ + Xγ)] – E₀ , where E_max is the maximum reduction in the log number CFU/thigh or ml, X is the PK-PD parameter being examined, EC₅₀ is the X value corresponding to 50% of the E_max, E₀ is the effect when X is 0, and γ is a sigmoidicity factor.