In Vitro Activity of Ceftazidime-Avibactam against Isolates in a Phase 3 Open-Label Clinical Trial for Complicated Intra-Abdominal and Urinary Tract Infections Caused by Ceftazidime-Nonsusceptible Gram-Negative Pathogens

ABSTRACT

The in vitro activity of ceftazidime-avibactam was evaluated against 341 Gram-negative isolates from 333 patients in a randomized, phase 3 clinical trial of patients with complicated urinary tract or intra-abdominal infections caused by ceftazidime-nonsusceptible pathogens (NCT01644643). Ceftazidime-avibactam MIC₉₀ values against Enterobacteriaceae and Pseudomonas aeruginosa (including several class B or D enzyme producers that avibactam does not inhibit) were 1 and 64 μg/ml, respectively. Overall, the ceftazidime-avibactam activity against ceftazidime-nonsusceptible isolates was comparable to the activity of ceftazidime-avibactam previously reported against ceftazidime-susceptible isolates. (This study has been registered at ClinicalTrials.gov under identifier NCT01644643.)

TEXT

Avibactam is the first in a class of new non-β-lactam β-lactamase inhibitors with a broader spectrum of inhibitory activity than previous generations of inhibitors against Ambler class A and C β-lactamases and some Ambler class D enzymes, including enzymes such as Klebsiella pneumoniae carbapenemase (KPC) and the carbapenem-hydrolyzing oxacillinase OXA-48 but with no activity against class B metallo-β-lactamases (1, 2). In combination, ceftazidime and avibactam have a spectrum of activity that includes extended-spectrum β-lactamase (ESBL)-producing Enterobacteriaceae, derepressed AmpC-positive Enterobacteriaceae, Pseudomonas aeruginosa, and isolates expressing serine carbapenemases such as KPC or OXA-48 (3, 4).

An open-label, phase 3 study (The REPRISE study [ClinicalTrials registration no. NCT01644643]) was conducted to compare the safety and efficacy of ceftazidime-avibactam (CAZ-AVI) with that of the best available therapy (∼97% received a carbapenem; the majority received this as monotherapy), as determined and documented prior to randomization by the treating physician, in complicated urinary tract infections (cUTI) and complicated intra-abdominal infections (cIAI) caused by Gram-negative pathogens nonsusceptible to ceftazidime (5). Patients were eligible for entry into the trial after the identification of at least one ceftazidime-nonsusceptible Gram-negative pathogen isolated from the site of infection. If the pathogen was susceptible to prior antibiotics, the protocol required that patients have either worsening signs and symptoms of cIAI/cUTI or a lack of improvement to be eligible for study entry. Specimens obtained from patients were processed at the local (or regional) laboratory according to local practices. Bacterial cultures were isolated from patient specimens and submitted to a central laboratory (Covance CLS, Indianapolis, IN) for identification and susceptibility testing. Susceptibility testing was performed using broth microdilution and interpreted according to Clinical and Laboratory Standards Institute (CLSI) methodologies (6, 7). FDA interpretive criteria were used for tigecycline (8). For ceftazidime-avibactam susceptibility testing, avibactam was tested at a constant concentration of 4 μg/ml in doubling dilutions of ceftazidime. All agents were tested by reference broth microdilution methods using frozen panels according to the manufacturer's recommendations (Trek Diagnostics, Westlake, OH). Phenotypic detection of ESBL enzyme production was performed according to the CLSI guidelines using the screening and confirmatory tests (6). Reference antibiotics included representative agents in relevant classes for comparative purposes. Genetic identification of common β-lactamases and upregulation of AmpC were provided as previously described (JMI Laboratories, Inc., North Liberty, IA) (9). Only baseline pathogens from all randomized patients were included in the analysis of susceptibility testing to ceftazidime-avibactam and comparators in this report.

In total, 341 baseline isolates of Enterobacteriaceae (314 isolates) and P. aeruginosa (27 isolates) from 333 randomized patients were obtained. There were 27 patients with cIAI and 306 patients with cUTI randomized in the study. Of the Enterobacteriaceae, Escherichia coli (139 isolates) was the most common organism isolated, followed by K. pneumoniae (131 isolates) and Enterobacter cloacae (17 isolates). The analysis included 31 isolates of Enterobacteriaceae that were susceptible to ceftazidime; however, these came from patients coinfected with a ceftazidime-nonsusceptible isolate. Susceptibilities to ceftazidime-avibactam and comparator agents are shown in Table 1. Ceftazidime-avibactam was very active against all of the Enterobacteriaceae, with overall MIC₅₀ and MIC₉₀ values of 0.25 and 1 μg/ml, respectively. The MIC distributions of ceftazidime and ceftazidime-avibactam are graphically depicted in Fig. 1. The ceftazidime-avibactam MIC distribution was shown to be ≤8 μg/ml for all isolates except four that produced a class B metallo-β-lactamase (which is not inhibited by avibactam). As expected, ceftazidime-avibactam MICs were high (≥32 μg/ml) for these four isolates, which were K. pneumoniae (2 isolates, 1 with NDM-1 and 1 with VIM-1), E. cloacae (1 isolate, with NDM-1), and Providencia rettgeri (1 isolate, with NDM-1). Of the remaining isolates, for which ceftazidime-avibactam MICs were ≤8 μg/ml, the majority possessed CTX-M and were usually associated with OXA-1/30 or other enzymes such as SHV-12. There were also 9 isolates that possessed the carbapenemases KPC (6 isolates of K. pneumoniae, 1 from a cIAI patient and 5 from cUTI patients) or OXA-48 (3 isolates of K. pneumoniae, all from cUTI patients). The ceftazidime-avibactam MICs ranged from 0.5 to 4 μg/ml (10). The MICs of ceftazidime alone were ≥64 μg/ml for all of these carbapenemase producers. Against 27 isolates of P. aeruginosa, the ceftazidime MIC₅₀ and MIC₉₀ values were 64 μg/ml and >64 μg/ml, and the ceftazidime-avibactam values were 8 μg/ml and 64 μg/ml, respectively. Of the 27 P. aeruginosa isolates, 13 possessed either a class B enzyme (1 isolate with VIM-2) or at least one class D enzyme (12 isolates with OXA-2, OXA-10, OXA-17, and/or OXA-74), which most likely contributed to the higher MIC₉₀ values observed in this analysis relative to those reported in previous studies against P. aeruginosa (11–13). The ceftazidime-avibactam MIC₉₀ value was 8 μg/ml for the 14 isolates without a class B or class D enzyme.

TABLE 1.

In vitro activity of ceftazidime-avibactam and other agents against Gram-negative isolates from patients enrolled in the phase 3 clinical trial (all randomized patients)^a

Baseline pathogen(s)	Agent(s)	No. of pathogens tested	MIC datab
			Range (μg/ml)	MIC50 (μg/ml)	MIC90 (μg/ml)	%Sc
Enterobacteriaceae
All pathogens	Ceftazidime-avibactam	314	≤0.008–>256	0.25	1	98.7
	Ceftazidime	314	0.06–>64	64	>64	9.9
	Amikacin	314	0.25–>64	4	>64	86.6
	Ciprofloxacin	314	0.015–>4	>4	>4	10.8
	Colistin	314	0.12–>128	1	2	91.4d
	Gentamicin	314	≤0.12–>16	>16	>16	40.1
	Imipenem	314	0.06–64	0.12	0.5	93.9
	Meropenem	314	0.015–>8	0.03	0.25	95.9
	Piperacillin-tazobactam	314	0.12–>128	16	>128	52.5
	Tigecycline	314	0.12–8	0.5	2	81.8e
	Trimethoprim-sulfamethoxazole	314	≤0.25–>8	>8	>8	34.7
Enterobacter cloacae	Ceftazidime-avibactam	17	0.25–>256	0.5	4	94.1
	Ceftazidime	17	16–>64	64	>64	0
	Amikacin	17	1–>64	2	>64	76.5
	Ciprofloxacin	17	0.015–>4	>4	>4	41.2
	Colistin	17	0.5–>128	1	2	94.1
	Gentamicin	17	0.25–>16	>16	>16	41.2
	Imipenem	17	0.12–16	0.25	0.5	94.1
	Meropenem	17	0.03–>8	0.06	0.25	94.1
	Piperacillin-tazobactam	17	2–>128	128	>128	23.5
	Tigecycline	17	0.25–8	1	4	58.8
	Trimethoprim-sulfamethoxazole	17	≤0.25–>8	≤0.25	>8	70.6
Escherichia coli	Ceftazidime-avibactam	139	≤0.008–8	0.12	1	100
	Ceftazidime	139	0.12–>64	32	>64	13.7
	Amikacin	139	0.5–>64	4	8	95.7
	Ciprofloxacin	139	0.015–>4	>4	>4	7.2
	Colistin	139	0.12–4	1	2	97.1
	Gentamicin	139	0.25–>16	16	>16	48.2
	Imipenem	139	0.06–1	0.12	0.25	100
	Meropenem	139	0.015–1	0.03	0.06	100
	Piperacillin-tazobactam	139	1–>128	8	>128	75.5
	Tigecycline	139	0.12–1	0.25	0.5	100
	Trimethoprim-sulfamethoxazole	139	≤0.25–>8	>8	>8	62.6
Klebsiella pneumoniae	Ceftazidime-avibactam	131	0.06–>256	0.5	1	98.5
	Ceftazidime	131	0.06–>64	>64	>64	5.3
	Amikacin	131	0.25–>64	4	>64	81.0
	Ciprofloxacin	131	0.03–>4	>4	>4	3.8
	Colistin	131	0.25–64	1	2	94.0
	Gentamicin	131	≤0.12–>16	>16	>16	32.1
	Imipenem	131	0.06–64	0.12	1	90.8
	Meropenem	131	0.015–>8	0.06	1	91.6
	Piperacillin-tazobactam	131	1–>128	128	>128	31.3
	Tigecycline	131	0.25–8	1	2	74.0
	Trimethoprim-sulfamethoxazole	131	≤0.25–>8	>8	>8	27.5
Other Enterobacteriaceaef	Ceftazidime-avibactam	27	0.03–256	0.5	2	96.3
	Ceftazidime	27	0.12–>64	64	>64	18.5
	Amikacin	27	0.5–>64	4	>64	74.1
	Ciprofloxacin	27	0.015–>4	>4	>4	18.5
	Colistin	27	0.5–>128	>128	>128	48.1d
	Gentamicin	27	0.25–>16	>16	>16	37.0
	Imipenem	27	0.12–16	0.5	4	77.8
	Meropenem	27	0.015–>8	0.06	0.25	96.3
	Piperacillin-tazobactam	27	0.12–>128	16	>128	55.6
	Tigecycline	27	0.25–8	2	8	40.7e
	Trimethoprim-sulfamethoxazole	27	≤0.25–>8	>8	>8	33.3
Nonfermenters
Pseudomonas aeruginosa	Ceftazidime-avibactam	27	1–256	8	64	55.6
	Ceftazidime	27	1–>64	64	>64	29.6
	Amikacin	27	0.5–>64	16	64	55.6
	Cefepime	27	1–>16	16	>16	29.6
	Ciprofloxacin	27	0.06–>4	>4	>4	18.5
	Colistin	27	0.5–4	2	4	100
	Gentamicin	27	1–>16	>16	>16	33.3
	Imipenem	27	0.5–>64	2	16	55.6
	Meropenem	27	0.06–>8	1	>8	59.3
	Piperacillin-tazobactam	27	0.5–>128	64	>128	33.3

^aMIC₅₀ and MIC₉₀ values were not calculated for pathogens with <10 patients. A patient can have more than one pathogen. Multiple isolates of the same species from the same patient are counted only once using the isolate with the highest MIC to the study drug received. For bacteremic patients, multiple isolates of the same species from the same patient are counted only once using the isolate with the highest MIC to the study drug received across the culture source (urine or blood).

^bThe total number of patients in the treatment group was 333.

^c%S, percent susceptible.

^dThe %S is based on EUCAST breakpoints and includes species with intrinsic resistance to colistin.

^eThe %S is based on EUCAST breakpoints and includes species with reduced susceptibility to tigecycline.

^fOther Enterobacteriaceae include Citrobacter freundii (7), Enterobacter aerogenes (2), Klebsiella oxytoca (1), Morganella morganii (15), Proteus mirabilis (16), Providencia rettgeri (15), Providencia stuartii (15), Raoultella terrigena (15), and Serratia marcescens (1).

FIG 1.

MIC frequency distribution of ceftazidime-avibactam and ceftazidime against ceftazidime-resistant Enterobacteriaceae isolated from patients enrolled in the phase 3 clinical trial.

The CLSI-recommended phenotypic test involving reduction of cephalosporin MICs with clavulanic acid showed that 216 Enterobacteriaceae isolates were phenotypically positive for an ESBL (Table 2). The ceftazidime-avibactam MIC₉₀ values for E. coli and K. pneumoniae were 0.5 and 1 μg/ml, respectively. In contrast, the ESBL phenotype-negative isolates of E. coli and K. pneumoniae were associated with higher ceftazidime-avibactam MIC₉₀ values of 8 and 4 μg/ml, respectively. This observation was most likely due to the presence of class B, C, or D β-lactamases that raised the MICs of ceftazidime but are not reduced by clavulanic acid in the ESBL tests. Organisms producing class B and some class D enzymes, which mask the ESBL test, are not inhibited by avibactam, thus resulting in the higher ceftazidime-avibactam MIC₉₀ values. This highlights a shortcoming of using only phenotypic tests to determine the presence or absence of an ESBL in isolates that carry multiple enzymes of different molecular classes.

TABLE 2.

In vitro activity of ceftazidime-avibactam against Gram-negative isolates from patients enrolled in the phase 3 clinical trial by ESBL phenotype (all randomized patients)^a

Baseline pathogen	ESBL statusb	No. of pathogens tested	MIC datac
			Range (μg/ml)	MIC50 (μg/ml)	MIC90 (μg/ml)	%Sd
Enterobacteriaceae
Escherichia coli	ESBL phenotype positive	108	≤0.008–2	0.12	0.5	100
	ESBL phenotype negative	31	0.06–8	0.25	8	100
Klebsiella pneumoniae	ESBL phenotype positive	106	0.12–2	0.5	1	100
	ESBL phenotype negative	25	0.06–>256	0.5	4	92.0
Proteus mirabilis	ESBL phenotype positive	2	0.06–0.5	NAe	NA	100
	ESBL phenotype negative	7	0.03–2	NA	NA	100

^aA patient could have more than one pathogen. Multiple isolates of the same species from the same patient are counted only once using the isolate with the highest MIC to the study drug received. For bacteremic patients, multiple isolates of the same species from the same patient are counted only once using the isolate with the highest MIC to the study drug received across the culture source (urine or blood).

^bGenetic identification of β-lactamases was provided as previously described (JMI Laboratories, Inc., North Liberty, IA) (9).

^cThe total number of patients in the treatment group was 333.

^d%S, percent susceptible.

^eNA, not applicable. MIC₅₀ and MIC₉₀ values were not calculated for pathogens with <10 patients.

Twenty-five baseline clinical isolates (Enterobacter spp., 13 isolates; P. aeruginosa, 7 isolates; Citrobacter freundii, 5 isolates) were hyperproducers of AmpC β-lactamase, either as a sole mechanism of resistance (10 isolates) or in combination with other enzymes (15 isolates). Of the 18 Enterobacteriaceae, only 5 isolates (2 E. cloacae, 2 Enterobacter aerogenes, and 1 C. freundii) had an upregulated AmpC as the sole mechanism of resistance. The ceftazidime MICs for 4 of these strains were ≥64 μg/ml, while the ceftazidime-avibactam MICs ranged from 0.25 to 1 μg/ml. One isolate of E. aerogenes, shown to be positive for upregulation of AmpC, was fully susceptible to ceftazidime. The ceftazidime MIC was 1 μg/ml, whereas the ceftazidime-avibactam MIC was 0.12 μg/ml. Of the 13 Enterobacteriaceae with upregulated AmpC in combination with other enzymes, the ceftazidime MICs ranged from 8 to ≥64 μg/ml, whereas the ceftazidime-avibactam MICs ranged from 0.12 to 4 μg/ml. Five of the seven isolates of P. aeruginosa were hyperproducers of AmpC β-lactamase as the sole β-lactamase responsible for ceftazidime resistance, whereas in two strains it was produced in combination with other acquired enzymes. Regardless of whether AmpC hyperproduction was the sole mechanism for ceftazidime resistance, there was a 2- to 4-dilution decrease in the MICs for ceftazidime-avibactam relative to those for ceftazidime, except in one strain in which there was only a one-dilution decrease in the MIC, from 64 μg/ml for ceftazidime to 32 μg/ml for ceftazidime-avibactam. This strain also possessed an OXA-17. Although no enzymatic inhibitory studies have been performed with avibactam and OXA-17, avibactam has been shown to be variable at inhibiting class D β-lactamases. The MIC for ceftazidime ranged from 16 to >64 μg/ml, whereas the range for ceftazidime-avibactam was 4 to 32 μg/ml.

In conclusion, ceftazidime-avibactam was highly active against isolates with the majority of MICs being ≤8 μg/ml in this phase 3 clinical study. This included isolates that were not susceptible to ceftazidime, due to ESBL, carbapenemases (KPC or OXA-48), and/or upregulated AmpC production. In addition, isolates of Enterobacteriaceae that had ceftazidime-avibactam MICs of >8 μg/ml produced class B enzymes that are not inhibited by avibactam. The ceftazidime-avibactam MIC₉₀ against P. aeruginosa isolated in this clinical study, which included a high proportion of strains with class B or D enzymes that avibactam does not inhibit, was 64 μg/ml. Ceftazidime-avibactam was shown to have good efficacy and was well tolerated in this global trial (5), suggesting the utility of ceftazidime-avibactam in treating patients with cUTI or cIAI caused by Gram-negative pathogens nonsusceptible to ceftazidime but susceptible to ceftazidime-avibactam. Overall, the activity of ceftazidime-avibactam against ceftazidime-nonsusceptible isolates was comparable to the activity of ceftazidime-avibactam previously reported against ceftazidime-susceptible isolates (14).