Considerations in the Selection of Renal Dosage Adjustments for Patients with Serious Infections and Lessons Learned from the Development of Ceftazidime-Avibactam

An extensive clinical development program (comprising two phase 2 and five phase 3 trials) has demonstrated the efficacy and safety of ceftazidime-avibactam in the treatment of adults with complicated intra-abdominal infection (cIAI), complicated urinary tract infection (cUTI), and hospital-acquired pneumonia (HAP), including ventilator-associated pneumonia (VAP). During the phase 3 clinical program, updated population pharmacokinetic (PK) modeling and Monte Carlo simulations using clinical PK data supported modified ceftazidime-avibactam dosage adjustments for patients with moderate or severe renal impairment (comprising a 50% increase in total daily dose compared with the original dosage adjustments) to reduce the risk of subtherapeutic drug exposures in the event of rapidly improving renal function.

ABSTRACT

An extensive clinical development program (comprising two phase 2 and five phase 3 trials) has demonstrated the efficacy and safety of ceftazidime-avibactam in the treatment of adults with complicated intra-abdominal infection (cIAI), complicated urinary tract infection (cUTI), and hospital-acquired pneumonia (HAP), including ventilator-associated pneumonia (VAP). During the phase 3 clinical program, updated population pharmacokinetic (PK) modeling and Monte Carlo simulations using clinical PK data supported modified ceftazidime-avibactam dosage adjustments for patients with moderate or severe renal impairment (comprising a 50% increase in total daily dose compared with the original dosage adjustments) to reduce the risk of subtherapeutic drug exposures in the event of rapidly improving renal function. The modified dosage adjustments were included in the ceftazidime-avibactam labeling information at the time of initial approval and were subsequently evaluated in the final phase 3 trial (in patients with HAP, including VAP), providing supportive data for the approved U.S. and European ceftazidime-avibactam dosage regimens across renal function categories. This review describes the analyses supporting the ceftazidime-avibactam dosage adjustments for renal impairment and discusses the wider implications and benefits of using modeling and simulation to support dosage regimen optimization based on emerging clinical evidence.

INTRODUCTION

Ceftazidime-avibactam, a combination antimicrobial agent comprising ceftazidime (a broad-spectrum antipseudomonal cephalosporin) and avibactam (a non-β-lactam β-lactamase inhibitor), is approved in Europe and the United States for the treatment of adults with complicated intra-abdominal infection (cIAI), complicated urinary tract infection (cUTI), and hospital-acquired pneumonia (HAP), including ventilator-associated pneumonia (VAP), caused by susceptible Gram-negative bacteria (1, 2). In Europe, it is also approved for the treatment of other infections caused by aerobic Gram-negative organisms in adult patients with limited treatment options (1). An extensive adult clinical development program, including two phase 2 and five phase 3 trials, has demonstrated the efficacy and safety of ceftazidime-avibactam versus carbapenem comparators in patients with serious bacterial infections, including those caused by ceftazidime-nonsusceptible and multidrug-resistant Gram-negative bacteria (3–9).

The approved ceftazidime-avibactam dosage regimen for adults with normal renal function or mild renal impairment (estimated creatinine clearance [CL_CR] of >50 ml/min, calculated using the Cockcroft-Gault formula [10]) is 2.5 g (2 g of ceftazidime plus 0.5 g of avibactam) every 8 h (q8h) by 2-h intravenous (i.v.) infusion (1, 2). Both ceftazidime and avibactam are predominantly cleared by the kidneys, and they have similar pharmacokinetic (PK) profiles and protein binding (11–13); hence, dosage adjustments are required for patients with moderate to severe renal impairment (CL_CR of ≤50 ml/min) (1, 2). Initial protocol-specified ceftazidime-avibactam dosage adjustments for patients with moderate to severe renal impairment utilized in the phase 3 studies were based on ceftazidime monotherapy labeling information and were guided by early population PK modeling and simulation data (13–15). In the absence of ceftazidime PK data for patients with CL_CR of <50 ml/min, the early population PK model and analyses of probability of target attainment (PTA) relied on ceftazidime data from the published literature and the labeled dosage recommendations for ceftazidime monotherapy. During the phase 3 program, ceftazidime-avibactam dosage adjustments for patients with renal impairment were reevaluated based on emerging clinical data from the phase 3 cIAI trial (RECLAIM), and updated population PK modeling and simulations were used to develop modified dosage regimens for patients with moderate or severe renal impairment, such that the total daily dose of ceftazidime-avibactam was increased by 50%, to reduce the risk of underexposure in patients with rapidly improving renal function (16, 17). The modified dosage adjustments for patients with renal impairment were included in the ceftazidime-avibactam labeling information at the time of U.S. and European approval (1, 2) and were subsequently validated in the final phase 3 trial of ceftazidime-avibactam in patients with HAP/VAP (REPROVE).

This review examines how clinical data and population PK modeling and simulations were used to guide the initial phase 3 ceftazidime-avibactam dosage modifications for patients with renal impairment and subsequently to update and validate the approved ceftazidime-avibactam dosage regimens across renal function categories from severe renal impairment to augmented renal clearance (ARC) and across multiple indications, including cIAI, cUTI, and HAP/VAP. The robust data and experience from the ceftazidime-avibactam development program represent an important case study in both the utility and limitations of PK/pharmacodynamic (PD) modeling in dose selection and dose optimization for antibiotic drug development.

CEFTAZIDIME-AVIBACTAM CLINICAL PROGRAM: OVERVIEW OF STUDIES AND INITIAL PHASE 3 DOSE SELECTION

Overview of the clinical trial program.

The ceftazidime-avibactam adult phase 2/3 program included two phase 2 trials (one in cIAI and one in cUTI) and five phase 3 international multicenter randomized trials in patients with infections proven or suspected to be caused by Gram-negative pathogens. The phase 3 program, comprising two trials in cIAI (RECLAIM and RECLAIM 3), one in cUTI (RECAPTURE), one in cIAI and cUTI caused by resistant pathogens (REPRISE), and one in HAP, including VAP (REPROVE), was conducted between 2012 and 2016 and included 3,712 patients (Fig. 1; see also Table S1 in the supplemental material). The study designs, methods, and efficacy and safety results of each trial have been reported separately (3–9). Clinical and microbiological outcomes were assessed at protocol-specified visits (Table S1). Each phase 3 study allowed for enrollment of patients with renal impairment (protocol-specified lower limits of CL_CR for exclusion varied from <6 ml/min to <30 ml/min [Table S1]). Dosage regimens for both ceftazidime-avibactam and the carbapenem comparators were adjusted for patients with renal impairment. The overall primary efficacy results across each of the phase 3 trials, and the results by patients’ baseline renal function subgroup, are presented in Table 1.

FIG 1.

Overview of the ceftazidime-avibactam phase 3 trials and population PK model iterations. cIAI, complicated intra-abdominal infection; cUTI, complicated urinary tract infection; HAP, hospital-acquired pneumonia; VAP, ventilator-associated pneumonia.

TABLE 1.

Clinical outcomes by renal function category across the ceftazidime-avibactam phase 3 program (5–9, 24, 25)^a

Study and outcome	No. with outcome/no. of subjects (%)
	All renal function categories		CLCR > 50 ml/min		CLCR ≤ 50 ml/min: original dosage regimen		CLCR ≤ 50 ml/min: modified dosage regimen		Augmented renal function (CLCR ≥ 151 ml/min)
	Ceftazidime-avibactam	Comparator	Ceftazidime-avibactam	Comparator	Ceftazidime-avibactam	Comparator	Ceftazidime-avibactam	Comparator	Ceftazidime-avibactam	Comparator
RECLAIM 1 and 2 (cIAI)
Clinical cure at TOC: mMITT	337/413 (81.6)	349/410 (85.1)	322/379 (85.0)	321/373 (86.1)	14/31 (45.2)	26/35 (74.3)
Clinical cure at TOC: MITT	429/520 (82.5)	444/523 (84.9)	407/476 (85.5)	410/478 (85.8)	20/41 (48.8)	32/43 (74.4)
Clinical cure at TOC: CE	376/410 (91.7)	385/416 (92.5)	356/383 (93.0)	362/390 (92.8)	18/25 (72.0)	22/25 (88.0)
REPRISE (cIAI and cUTI)
Clinical cure at TOC: mMITT (cIAI)	8/10 (80.0)	6/11 (54.5)			0/0	3/5 (60.0)
Clinical cure at TOC: mMITT (cUTI)	132/144 (91.7)	129/137 (94.2)			24/26 (92.3)	22/24 (91.7)
RECAPTURE 1 and 2 (cUTI)
Day 5 symptomatic response: mMITT	276/393 (70.2)	276/417 (66.2)	245/350 (70.0)	252/379 (66.5)	30/42 (71.4)	24/35 (68.6)
Combined response at TOC: mMITT	280/393 (71.2)	269/417 (64.5)	251/350 (71.7)	244/379 (64.4)	28/42 (66.7)	25/35 (71.4)
Microresponse at TOC: mMITT	304/393 (77.4)	296/417 (71.0)	274/350 (78.3)	268/379 (70.7)	29/42 (69.0)	27/35 (77.1)
RECLAIM 3 (cIAI)
Clinical cure at TOC: CE	166/177 (93.8)	173/184 (94.0)	156/166 (94.0)	162/172 (94.2)	10/11 (90.9)	11/12 (91.7)
REPROVE (HAP, including VAP)b
Clinical cure at TOC: cMITT	245 (68.8)	270 (73.0)	187 (65.4)	208 (71.2)	18/30 (60.0)	16/28 (57.1)	13/18 (72.2)	12/18 (66.7)	43 (86.0)	48 (82.8)
Clinical cure at TOC: CE	199 (77.4)	211 (78.1)	156 (74.6)	165 (77.1)	15/21 (71.4)	13/23 (56.5)	11/13 (84.6)	12/17 (70.6)	31 (91.2)	33 (86.8)
28-day mortality: ITT	42/436 (9.6)	36/434 (8.3)	33/313 (10.5)	25/314 (8.0)	4/31 (12.9)	6/31 (19.4)	3/21 (14.3)	2/19 (10.5)	2/58 (3.4)	3/62 (4.8)

^aClinical outcomes shown are the primary/coprimary endpoint(s) defined by protocol for each respective study. CE, clinically evaluable; cIAI, complicated intra-abdominal infection; CL_CR, creatinine clearance; cMITT, clinically modified intention-to-treat; cUTI, complicated urinary tract infection; HAP, hospital-acquired pneumonia; ITT, intention-to-treat; mMITT, microbiologically modified intention-to-treat; TOC, test of cure; VAP, ventilator-associated pneumonia.

^bThe cMITT and CE primary efficacy analysis in REPROVE excluded patients with moderate or severe impairment (CL_CR of 16–50 ml/min) who received the original ceftazidime-avibactam dosage regimens for renal impairment; the ITT population included patients who received both the original and modified ceftazidime-avibactam dosage regimens (dosage adjustments for the comparator, meropenem, remained the same throughout the study).

Population PK modeling and simulations.

Population PK modeling of ceftazidime and avibactam was done iteratively throughout the development process as additional patient PK data became available from the clinical studies (Fig. 1), to predict concentration-time profiles and calculate steady-state PK parameters (such as maximum concentration at steady state [C_max,_ss] and area under the concentration-time curve from 0 to 24 h at steady state [AUC_ss]). Monte Carlo simulations were used in PTA analyses evaluating the percentage of simulated patients in each renal function category who achieved the PK/pharmacodynamic (PD) targets for both ceftazidime and avibactam simultaneously (joint PTA), defined as free plasma ceftazidime concentrations above the target ceftazidime-avibactam MIC of 8 mg/liter for at least 50% of the dosing interval (50% fT>MIC) and free plasma avibactam concentrations above a threshold concentration (C_T) of 1 mg/liter for at least 50% of the dosing interval (50% fT > 1 mg/liter) (18–20). A MIC of ≤8 mg/liter has been established as the susceptibility breakpoint for ceftazidime-avibactam against Enterobacteriaceae and Pseudomonas aeruginosa based on the approved dosage regimens (21).

An early iteration of the ceftazidime and avibactam models (designated the phase 2 model) was used to support initial phase 3 dose selection (14, 15). Later (interim phase 3 and final phase 3) model iterations (based on emerging clinical data) were used to select modified dosage adjustments for renal function and to validate model-predicted exposures across patient subgroups (16, 17). Six renal function categories were evaluated in the simulations based on CL_CR: normal (>80 ml/min), mild impairment (51 to 80 ml/min), moderate impairment (31 to 50 ml/min), severe impairment 1 (16 to 30 ml/min), severe impairment 2 (6 to 15 ml/min), and end-stage renal disease (ESRD; <6 ml/min). For patients with normal renal function, the distribution of patient-specific covariates was based on those observed in prior clinical studies. Within renal impairment categories, CL_CR values were assumed to follow a uniform distribution within the designated range for each category.

Initial ceftazidime-avibactam phase 3 dose selection.

Ceftazidime-avibactam dosage selection for the phase 3 program (Table 2) was based on extensive nonclinical, microbiological, and PK data for avibactam alone and in combination with ceftazidime, clinical data for ceftazidime alone, and the phase 2 population PK model, which incorporated ceftazidime PK data from 103 subjects and avibactam PK data from 288 subjects, from 11 phase 1 and two phase 2 trials (14, 15). Dosage adjustments for patients with CL_CR of ≤50 ml/min were based on the approved ceftazidime labeling information (22, 23), as well as data from the literature (11) and one phase 1 study (13), with achievement of joint PTA of >90% (rather than matching exposures to patients with normal renal function) considered the main criterion for dosage regimen selection; notably, the analyses assumed patients had stable renal function. These data demonstrated that the original dosage regimens were associated with predicted exposures and PTA values of >90% in all renal function categories except for severe renal impairment 1 (Table 3 and Table S2); however, comparison of the simulation results across different targets indicated that ceftazidime, and not avibactam, was the limiting factor for joint attainment (data not shown). Thus, given the established use of ceftazidime in patients with renal impairment, the proposed ceftazidime-avibactam dosage adjustments for renal impairment based on the ceftazidime labeling information (22, 23) were considered appropriate.

TABLE 2.

Original and modified ceftazidime-avibactam dosage regimens by renal function category based on estimated creatinine clearance calculated using the Cockcroft-Gault formula (10)^a

Renal function category (estimated CLCR, ml/min)	Original dosage regimen	Modified (labeled) dosage regimen
Normal (>80)	2,000 mg of ceftazidime + 500 mg of avibactam q8h	No change
Mild impairment (51–80)	2,000 mg of ceftazidime + 500 mg of avibactam q8h	No change
Moderate impairment (31–50)	1,000 mg of ceftazidime + 250 mg of avibactam q12h	1,000 mg of ceftazidime + 250 mg of avibactam q8h
Severe impairment 1 (16–30)	1,000 mg of ceftazidime + 250 mg of avibactam q24h	750 mg of ceftazidime + 187.5 mg of avibactam q12h
Severe impairment 2 (6–15)	500 mg of ceftazidime + 125 mg of avibactam q24h	750 mg of ceftazidime + 187.5 mg of avibactam q24h
End-stage renal disease (<6)	500 mg of ceftazidime + 125 mg of avibactam q48h	750 mg of ceftazidime + 187.5 mg of avibactam q48h

^aAll ceftazidime-avibactam doses were given as 2-h i.v. infusions. i.v., intravenous; q8h, every 8 h; q12h, every 12 h; q24h, every 24 h; q48h, every 48 h.

TABLE 3.

Joint PTA by renal function category and indication for simulated patients receiving original and modified ceftazidime-avibactam dosage regimens for renal impairment (16, 17, 27, 36)^a

Renal function category (CLCR, ml/min)	Joint PTA, %
	Original dosage regimens					Modified dosage regimens
	Phase 2 model, cIAI	Interim phase 3 model				Interim phase 3 model				Final phase 3 model
		cIAI	cUTI	NP	VAP	cIAI	cUTI	NP	VAP	cIAI	cUTI	NP	VAP
Normal (>80)	96.2	98.4	99.6	95.5	94.9	98.4	99.6	95.5	94.9	94.9	95.2	98.3	96.1
Mild renal impairment (51–80)	98.9	99.7	99.5	99.6	99.7	99.7	99.5	99.6	99.7	99.0	98.7	98.9	97.6
Moderate impairment (31–50)	95.8	97.0	97.3	95.9	96.3	99.7	99.9	99.7	99.7	99.3	99.1	98.8	97.7
Severe renal impairment 1 (16–30)	89.7	88.3	94.2	85.4	87.4	99.7	99.5	99.7	99.7	99.0	98.6	98.8	97.3
Severe impairment 2 (6–15)	95.7	97.1	99.1	96.8	97.0	99.8	99.7	99.7	99.8	99.3	98.8	99.2	98.0
End-stage renal disease (<6)	99.1	99.4	98.2	99.3	99.4	99.7	100.0	99.7	99.7	99.6	99.5	99.5	98.8

^aAll ceftazidime-avibactam doses were given as 2-h i.v. infusions. See Table 2 for original and modified dosage regimens.

CEFTAZIDIME-AVIBACTAM PHASE 3 CLINICAL OUTCOMES BY RENAL FUNCTION

Phase 3 cIAI study: RECLAIM.

In RECLAIM, patients with cIAI (n = 1,066) were randomized 1:1 to receive either ceftazidime-avibactam (2.5 g q8h for patients with CL_CR of >50 ml/min) plus metronidazole (0.5 g q8h) or meropenem (1 g q8h) for 5 to 14 days (5); patients with CL_CR of ≤50 ml/min received the original ceftazidime-avibactam dosage adjustments (Table 2) or the standard meropenem dosage adjustments as recommended in the product label. The results demonstrated overall noninferiority of ceftazidime-avibactam plus metronidazole versus meropenem for the primary endpoint of clinical cure rate at the test-of-cure (TOC) visit (Table 1), with the safety profile of ceftazidime-avibactam plus metronidazole in patients with cIAI as expected for ceftazidime alone. Among patients with normal renal function or mild renal impairment, clinical cure rates were similar between treatment groups. However, clinical cure rates in the subgroup of patients with moderate renal impairment at baseline (CL_CR of 31 to 50 ml/min) were lower in both treatment groups than in the overall study population and within this subgroup were disproportionately lower for ceftazidime-avibactam than for meropenem (Table 1) (5).

Of note, the subgroup of patients with moderate renal impairment at baseline was a limited subset (8% of patients), and caution in interpreting this finding is required; analysis of the patients with unfavorable outcomes suggested that the majority of such outcomes were unrelated to lack of efficacy of study drug. However, review of serum creatinine (SCr) levels revealed that rapid improvements in renal function in the postrandomization early treatment period (possibly due to fluid resuscitation and surgery) occurred among approximately 70% of patients with moderate renal impairment in both treatment groups. The decrease in response rates in this subset in both treatment groups may thus be attributable in part to rapidly improving CL_CR (within the first 72 h) in some patients and a corresponding potential mismatch between their actual CL_CR and the dose of study drug they were receiving based on their CL_CR at baseline, such that they may have been underdosed early in the course of treatment, a critical period that can determine clinical outcome (5). Although dosage adjustments were made in response to changes in SCr levels, the dynamic nature of CL_CR in patients with cIAI and the intermittent measurement of SCr in clinical practice mean that such mismatches are inevitable, even if adjustments are made promptly in response to new SCr measurements. Moreover, the protocol-specified adjustment in total daily dose for patients with moderate renal impairment treated with ceftazidime-avibactam (1.25 g q12h, an ∼66% reduction compared to standard dose) represented a greater reduction of the dose for normal or mild impairment than for meropenem (1 g q12h, an ∼33% reduction). Therefore, the risk of underdosing and consequential inadequate exposure was greater for ceftazidime-avibactam than for meropenem (5).

At the time that the RECLAIM data became available, the other ceftazidime-avibactam phase 3 studies were ongoing and at various stages of completion. Since the primary efficacy endpoints in RECLAIM had been met, there was confidence that the ceftazidime-avibactam dosage regimen for patients with CL_CR of >50 ml/min was adequate for the ongoing trials. Moreover, the available PK data indicated that all subjects had achieved adequate exposure for their baseline renal function category, and among patients for whom underdosing was likely not an issue (those with baseline CL_CR of >50 ml/min), the demonstration of efficacy validated the early models and PTA analyses. However, the findings prompted further modeling and simulations (described below), with the aim of developing modified dosage regimens for renal impairment to address the possibility of underdosing and ensure adequate drug exposure in the setting of rapid improvements of renal function early in the course of treatment. While these analyses were ongoing, the phase 3 Asia cIAI study (RECLAIM 3) and phase 3 HAP/VAP (REPROVE) study protocols were amended to exclude enrollment of patients with CL_CR of ≤50 ml/min, given that these studies were still recruiting patients.

Phase 3 resistant pathogen study: REPRISE.

REPRISE was an open-label, pathogen-directed, randomized study comparing ceftazidime-avibactam with best available therapy (BAT; 97% carbapenems) in patients with cIAI (n = 21) or cUTI (n = 281) caused by ceftazidime-nonsusceptible Enterobacteriaceae or P. aeruginosa (6). All patients received the standard approved ceftazidime-avibactam dose for estimated CL_CR of >50 ml/min and the original dosage adjustments for CL_CR of ≤50 ml/min (Table 2). Overall clinical cure rates at TOC were similar for ceftazidime-avibactam and BAT (Table 1), providing important evidence of the efficacy for ceftazidime-avibactam in the treatment of patients with serious infections caused by ceftazidime-nonsusceptible Enterobacteriaceae and P. aeruginosa (6). In the cIAI cohort, there were no patients in the ceftazidime-avibactam arm with baseline CL_CR of ≤50 ml/min. In the cUTI cohort, clinical cure rates at TOC were similar for ceftazidime-avibactam and BAT overall and were also similar between groups in the subset of patients with baseline CL_CR of ≤50 ml/min (Table 1).

Phase 3 cUTI study: RECAPTURE.

In RECAPTURE, patients with cUTI, including acute pyelonephritis (n = 1,033), were randomized 1:1 to receive ceftazidime-avibactam or doripenem, with an optional switch to oral therapy after 5 days of i.v. study treatment (total i.v. ± oral treatment duration of 10 to 14 days) (7). As in RECLAIM and REPRISE, all patients received the standard dose and original ceftazidime-avibactam dosage adjustments for estimated CL_CR of >50 ml/min and for CL_CR of ≤50 ml/min, respectively (Table 2). Ceftazidime-avibactam demonstrated noninferiority for the FDA coprimary endpoints of (i) patient-reported symptomatic resolution at day 5 and (ii) combined symptomatic resolution/microbiological eradication at TOC, with the difference in microbiological eradication rates at TOC suggesting statistical superiority in favor of ceftazidime-avibactam (7). There was no evidence of differences in any of the primary outcome measures for patients with baseline CL_CR of ≤50 ml/min versus those with CL_CR of >50 ml/min (Table 1).

Phase 3 cIAI ASIA study: RECLAIM 3.

RECLAIM 3 was similar in design to RECLAIM but was conducted solely in Asian countries (8). Patients with cIAI (n = 441) were randomized 1:1 to receive ceftazidime-avibactam plus metronidazole or meropenem for 5 to 14 days. Patients received the standard approved ceftazidime-avibactam dose for estimated CL_CR of >50 ml/min and the original dosage adjustments for CL_CR of ≤50 ml/min (Table 2), with the protocol amended partway through the study to exclude patients with CL_CR of ≤50 ml/min. The primary endpoint was met, with ceftazidime-avibactam plus metronidazole demonstrating noninferiority to meropenem for the difference in clinical cure rates at TOC (Table 1). Clinical cure rates at TOC in the limited subgroup of patients with CL_CR of 30 to 50 ml/min were similar to those in the overall study population (Table 1).

Phase 3 HAP/VAP study: REPROVE.

REPROVE, a double-blind noninferiority trial comparing ceftazidime-avibactam with meropenem in the treatment of patients with HAP, including VAP, was the final study in the adult phase 3 program (9, 24). Overall, 879 patients with HAP (including VAP) were randomized 1:1 to receive ceftazidime-avibactam or meropenem for 7 to 14 days. All patients with estimated CL_CR of >50 ml/min received the standard ceftazidime-avibactam dose. As REPROVE was ongoing at the time the RECLAIM results became available, the study protocol was amended, initially to exclude patients with baseline CL_CR of ≤50 ml/min and subsequently to implement the modified ceftazidime-avibactam dosage regimens for CL_CR of ≤50 ml/min, which are aligned with the current product labeling (1, 2). In total, 62 patients with moderate or severe renal impairment at baseline were recruited prior to the protocol amendments, and a further 40 patients with moderate or severe renal impairment were enrolled after the modified dosages were implemented; results (including primary efficacy outcomes) for these patients were summarized separately (9, 24, 25). Overall, ceftazidime-avibactam demonstrated noninferiority to meropenem against the protocol-specified primary endpoint of clinical cure at TOC and an FDA-guided endpoint of 28-day mortality (Table 1).

As REPROVE included HAP and VAP patients with moderate or severe renal impairment at baseline who received either the original or modified ceftazidime-avibactam dosage adjustments, the data provide an opportunity to compare outcomes for these respective patient subgroups. Again, with the important caveat that the small numbers limit interpretation, clinical cure rates at TOC, as well as 28-day mortality rates in patients with moderate or severe renal impairment who received the original ceftazidime-avibactam dosage regimens prior to the protocol amendments, were similar to those for the overall study population, as well as to those in patients who received the modified ceftazidime-avibactam dosage regimens for CL_CR of ≤50 ml/min (Table 1). These findings are consistent with the other phase 3 ceftazidime-avibactam trial outcomes (with the exception of RECLAIM), suggesting that the original dosage adjustment for moderate renal impairment provided adequate exposure in those studies.

CEFTAZIDIME-AVIBACTAM DOSE OPTIMIZATION AND VALIDATION

Interim phase 3 population PK models.

Following the results indicating potential underdosing of patients with moderate renal impairment in RECLAIM, the population PK models were updated (designated the interim phase 3 model) with PK data from RECLAIM and the first data cutoff from REPRISE and included data for ceftazidime and avibactam from 780 and 1,044 subjects, respectively (16, 26). Although unblinded PK data for patients with HAP/VAP were not available for inclusion in these models, blinded ceftazidime and avibactam concentration data from 78 HAP/VAP patients in REPROVE were compared to data from cIAI patients at equivalent sampling windows. The observed concentrations were similar, so simulations for patients with nosocomial pneumonia (NP)/VAP were based on the models in patients with cIAI and incorporated blinded demographics data from the (then ongoing) REPROVE study. These models were used to predict exposures and joint PTA for patients with cIAI, cUTI, NP, or VAP receiving the original phase 3 dosages and various potential modified dosage regimens for each renal function category (16, 26). To assess the effect of rapid improvements in renal function, ceftazidime and avibactam exposures and joint PTA were also calculated for patients receiving the recommended (original and modified) dosage regimens for renal function one category below (i.e., more severe impairment) than the actual current renal status of the patient and for patients with normal renal function (CL_CR of >80 ml/min) receiving dosage regimens for patients with moderate renal impairment (CL_CR of 30 to 50 ml/min) (16, 26).

Reevaluation of original dosage adjustments for patients with renal impairment.

(i) Exposure and PTA simulations. Based on the interim ceftazidime and avibactam phase 3 population PK models, simulations demonstrated that the original dosage adjustments for the different renal function categories and indications achieved >90% joint PTA, except for patients with cIAI, NP, or VAP and severe renal impairment 1 (CL_CR of 16 to 30 ml/min), for whom PTA was approximately 85 to 88% (Table 3 and Table S3). Ceftazidime and avibactam exposures in simulated cIAI patients (geometric mean C_max,ss and AUC_ss) with moderate renal impairment (CL_CR of 31 to 50 ml/min), severe impairment 1 (CL_CR of 16 to 30 ml/min), or severe impairment 2 (CL_CR of 6 to 15 ml/min) were lower than those in cIAI patients with normal renal function, except in patients with ESRD (CL_CR of <6 ml/min), for whom geometric mean AUC_ss values were approximately 50% higher than in cIAI patients with normal renal function but comparable to those in cIAI patients with mild renal impairment (27).

(ii) Rapidly improving renal function. Simulations also assessed the hypothesis that patients with rapidly improving renal function without a timely change in dosage were underdosed. PTAs for the original dosage regimens for moderate impairment, severe impairment 1, severe impairment 2, and ESRD in the event of renal function improvement to the next highest renal function category were low across indications, ranging from 23.6% for NP to 68.2% for cUTI (Table 4). PTAs for the original dosage regimens given to patients with moderate renal impairment in the event of renal function improvement to normal renal function were well below the targeted range (12.1 to 22.7% across indications [Table 4]).

TABLE 4.

Joint PTA by renal function category and indication for 5,000 simulated patients treated with original and modified ceftazidime-avibactam dosage regimens for renal impairment when renal function is at the next immediately higher category^a (interim phase 3 population PK model) (16, 27)

Renal function category (CLCR, ml/min)		Dosage regimen	Joint PTA, %
Actual	Treated		cIAI	cUTI	NP	VAP
Normal (>80)	Moderate impairment (31–50)	Original: 1,000 mg of ceftazidime + 250 mg of avibactam, q12h	13.5	22.7	12.7	12.1
		Modified: 1,000 mg of ceftazidime + 250 mg of avibactam, q8h	55.9	82.0	52.5	50.3
Mild impairment (51–80)	Moderate impairment (31–50)	Original: 1,000 mg of ceftazidime + 250 mg of avibactam, q12h	52.5	68.2	49.3	52.0
		Modified: 1,000 mg of ceftazidime + 250 mg of avibactam, q8h	95.7	97.0	94.2	95.1
Moderate impairment (31–50)	Severe impairment 1 (16–30)	Original: 1,000 mg of ceftazidime + 250 mg of avibactam q24h	23.9	37.6	23.6	25.0
		Modified: 750 mg of ceftazidime + 187.5 mg of avibactam, q12h	85.5	95.0	83.2	84.8
Severe impairment 1 (16–30)	Severe impairment 2 (6–15)	Original: 500 mg of ceftazidime + 125 mg of avibactam, q24h	32.0	52.9	31.4	32.5
		Modified: 750 mg of ceftazidime + 187.5 mg of avibactam, q24h	68.8	83.0	66.4	68.6
Severe impairment 2 (6–15)	ESRD (<6)	Original: 500 mg of ceftazidime + 125 mg of avibactam, q48h	44.5	63.2	43.6	44.5
		Modified: 750 mg of ceftazidime + 187.5 mg of avibactam, q48h	75.2	86.2	73.7	75.2

^aThe dose for patients with moderate renal impairment was evaluated for both the mild impairment and normal renal function groups. All ceftazidime-avibactam doses were given as 2-h i.v. infusions. See Table 2 for original and modified dosage regimens.

Development of modified dosage adjustments for patients with renal impairment.

(i) Overall approach. To mitigate the potential impact of ceftazidime-avibactam underdosing in the event of rapid improvement in renal function, PTA analyses using the interim phase 3 models explored different dosage regimens for the moderate impairment, severe impairment 1, severe impairment 2, and ESRD renal function groups (16). Because the predicted ceftazidime and avibactam exposures for the original dosage adjustments were generally lower for patients with moderate to severe renal impairment than in patients with normal renal function receiving the standard dose, there was some flexibility to increase the dose and/or administration frequency. The aim was to identify modified dosage regimens that would provide improved joint PTA in cases in which a patient with rapidly improving renal function might be receiving the recommended dosage regimen for a renal impairment category one grade more severe than their actual current renal status, without exceeding observed exposures in patients with normal renal function or mild renal impairment receiving the standard dose. Consistent with ceftazidime labeling information (22, 23), no dosage adjustments were considered for patients with mild renal impairment (CL_CR of >50 ml/min). For each renal function group (other than normal and mild impairment, as noted above), various possible ceftazidime-avibactam dosage regimens were explored which either increased the total daily dose by 50% with the same or lower dose unit but more frequent infusions or increased the dose amount by 50% while maintaining the same administration frequency as in the original regimens. The modified dosage regimens that were ultimately selected (Table 2) were evaluated based on predefined criteria, including exposures comparable to or lower than in and joint PTAs comparable to or higher than in adults with normal renal function receiving the standard dosage regimen (27).

(ii) Exposure and PTA simulations. PTAs for the original and modified dosage regimens are shown in Table 3; ceftazidime and avibactam exposures (geometric mean C_max,ss and AUC_ss) are shown in Tables S3 and S4, respectively. For simulated patients with moderate renal impairment, the modified ceftazidime-avibactam dosage regimen (1,000 mg of ceftazidime plus 250 mg of avibactam q8h) resulted in ceftazidime and avibactam exposures that were slightly higher than in patients with normal renal function but lower than in patients with mild impairment receiving the standard dose (Table S4). For simulated patients with severe renal impairment 1 or severe renal impairment 2, the modified ceftazidime-avibactam dosage regimens (750 mg of ceftazidime plus 187.5 mg of avibactam q12h and 750 mg of ceftazidime plus 187.5 mg of avibactam q24h, respectively) resulted in geometric mean ceftazidime and avibactam AUC_ss values that were comparable to those in patients with normal renal function receiving the standard dose; respective C_max,ss values were slightly lower and minimum concentration at steady state (C_min,ss) values were slightly higher than in patients with normal renal function receiving the standard dose. For simulated patients with ESRD, the selected modified ceftazidime-avibactam dosage regimen (750 mg of ceftazidime plus 187.5 mg of avibactam q48h) resulted in geometric mean ceftazidime AUC_ss values that were approximately 48% higher, and avibactam AUC_ss values that were approximately 25% lower, than in patients with mild renal impairment receiving the standard dose (this group was selected for comparison because exposures were higher than for patients with normal renal function but still considered within acceptable limits for safety); respective C_max,ss values were slightly lower than, or comparable to, those in patients with mild renal impairment receiving the standard dose (Table S4). An increase in AUC_ss of less than 50% for patients with ESRD was considered acceptable because such patients are generally under hemodialysis 2 or 3 times per week and part of the accumulated ceftazidime and avibactam will be cleared during hemodialysis (27).

(iii) Rapidly improving renal function. As with the original dosage adjustments, simulations also assessed joint PTA in the event of renal function improvement to the next highest renal function category for the modified ceftazidime-avibactam dosage adjustments, and also for the modified moderate renal impairment regimen in patients with normal renal function (Table 4). While it was not possible to increase exposures in the case of a rapid improvement in renal function to match those in patients with normal function or mild impairment receiving the standard dose, joint PTA did improve from approximately 13 to 50% for the original regimens to approximately 50 to 95% for the modified dosages depending on indication and renal function group (Table 4).

Validation of the modified dosage regimens: final phase 3 population PK models.

After completion of the phase 3 program, exposure and PTA analyses similar to those described above were conducted using updated phase 3 population PK models that included all previous model data plus patient PK data from RECLAIM 3, REPRISE, RECAPTURE, and REPROVE, including ceftazidime and avibactam data from 1,975 and 2,249 subjects, respectively, designated the final phase 3 model (17). These analyses confirmed that the selected modified dosage regimens for each renal function category provided appropriate ceftazidime and avibactam exposures in simulated patients (Table S5) and high PTAs (>90%) across indications (Table 1) (17). Individual predicted exposures and joint target attainment among phase 3 trial patients with PK data available for analysis (Table 5) were also >95%, except for the ESRD group, reflecting the small number of patients with ESRD enrolled across the trials.

TABLE 5.

Individual predicted steady-state exposures and joint PK/PD target attainment for ceftazidime and avibactam in phase 3 patients by baseline renal function (final phase 3 population PK model) (17)^a

CLCR, ml/min	n	Ceftazidime		Avibactam		Joint target attainment, % (95% CI)
		Cmax,ss, mg/liter	AUCss, mg·h/liter	Cmax,ss, mg/liter	AUCss, mg·h/liter
8–15	4	34.3 (173.3)	551 (121.9)	6.3 (305.6)	86 (220.6)	75.0 (32.6, 100.0)
>15–30	20	50.4 (139.5)	789 (116.5)	10.9 (174.1)	155 (143.6)	100.0 (NA)
>30–50	128	58.8 (120.5)	938 (122.9)	10.2 (147.6)	148 (153.3)	98.4 (96.3, 100.0)
>50–80	418	90.0 (108.0)	1,213 (110.4)	15.3 (142.9)	186 (144.5)	99.0 (98.1, 100.0)
>80–150	955	72.9 (105.9)	828 (112.4)	13.2 (165.5)	138 (163.4)	99.0 (98.3, 99.6)
>150–180	123	58.5 (93.0)	652 (112.8)	9.9 (124.5)	103 (137.5)	98.4 (96.1, 100.0)
>180–610	116	51.2 (109.6)	542 (108.1)	9.9 (171.6)	96 (155.9)	95.7 (92.0, 99.4)

^aAll ceftazidime-avibactam doses were given as 2-h i.v. infusions. Exposure values are geometric means, with percent coefficient of variation in parentheses. AUC_ss, area under the concentration-time curve from 0 to 24 h at steady state; CI, confidence interval; C_max,ss, maximum plasma concentration at steady state; NA, not applicable.

Augmented renal clearance.

As patients who are critically ill frequently experience ARC (defined here as estimated CL_CR of >150 ml/min), it is important to evaluate the effect of high CL_CR on antibiotics and other renally cleared drugs that may be used in this setting. Although each of the ceftazidime-avibactam phase 3 trials included patients with potential ARC at baseline, REPROVE was the only trial to include prespecified outcomes analyses for this patient group. In the subset of the REPROVE patient population with potential ARC, clinical cure and 28-day mortality rates among patients who received the standard ceftazidime-avibactam dosage regimen for patients with CL_CR of >50 ml/min were similar to those in patients who received meropenem (Table 1). To assess any effect of potential ARC (defined as CL_CR of >150 to 180 ml/min and >180 to 610 ml/min) on ceftazidime and avibactam PK, individual predicted patient concentration-time profiles from the final population PK model were used to determine individual patient exposures and target attainment for patients with CL_CR in the normal range (>80 to 150 ml/min) and those with potential ARC (Table 5). Joint target attainment was maintained in >95% of patients with CL_CR of >150 ml/min, despite decreases in AUC_ss by up to 35% relative to those with normal renal function (17). The ability to provide adequate exposures for patients with ARC without increasing ceftazidime-avibactam doses beyond the standard dose for patients with normal renal function is explained by the population PK models, in which the relationships between CL_CR and CL for CL_CR values of >100 ml/min and >80 ml/min for ceftazidime and avibactam, respectively, become shallow, such that exposures only decrease modestly at high CL_CR values (17).

DISCUSSION

Providing adequate dosing for clinical efficacy and safety while adjusting for renal function is of critical importance for outcomes of antibiotic therapy, particularly for certain patient groups, such as the elderly, obese, or critically ill (28–30). As highlighted by other authors (31, 32), reduced efficacy in patients with renal impairment has been observed in phase 3 trials of several antibiotics, including ceftazidime-avibactam, ceftolozane-tazobactam, daptomycin, and telavancin, potentially as a result of improving renal function during therapy with consequent underdosing due to a reduced dose being given based on initial creatinine clearance assessment. Ensuring adequate exposures in critically ill patients, who frequently have ARC and other physiological changes that can alter antibiotic PK, is an important consideration for dose selection (28). The results from the ceftazidime-avibactam adult phase 3 program demonstrated the efficacy and safety of this combination across a range of serious infections but also highlight the need to carefully monitor renal function and adjust dosages promptly and appropriately, particularly early in the course of treatment (during which rapid improvements can occur), to ensure that exposures are adequate. This need for additional precaution and monitoring in patients with moderate or severe renal impairment is reflected in the ceftazidime-avibactam product labeling information (1, 2). A similar reduction in clinical cure rates among patients with CL_CR of ≤50 ml/min has been observed in phase 3 clinical trials comparing ceftolozane-tazobactam with meropenem in patients with cIAI, although dosing regimens for this combination have not been modified to account for rapidly improving renal function (33, 34). In addition, reduced efficacy in patients with CL_CR of ≤50 ml/min has been reported in phase 3 trials of daptomycin in skin and soft tissue infections (cSSTI) and bacteremia or right-sided endocarditis and for telavancin in patients with cSSTI or HAP (31); it is likely that similar issues will apply to other antibiotics that undergo renal clearance. While there may be factors unrelated to antibiotic dosing that contribute to reduced therapeutic responses in some patients with renal impairment, interest in this topic is growing, and various potential solutions have been proposed. Bidell and Lodise have suggested that other methods of calculating renal function that do not assume stable renal function, such as kinetic estimated glomerular filtration rate, should be evaluated for their potential to improve antibiotic dosing in practice, particularly in patients with acute renal impairment and potentially rapidly changing renal function (31). In view of the potential for rapid improvements in renal function in infected patients with acute kidney injury, and the critical importance of adequate exposures particularly during the early stages of treatment, an alternative approach to avoiding antibiotic underdosing (for drugs with a wide therapeutic index, such as β-lactams) suggested by Crass et al. involves deferment of dosage adjustments for the first 48 h of treatment (32).

In the pivotal phase 3 cIAI trial (RECLAIM), ceftazidime-avibactam plus metronidazole demonstrated noninferiority versus meropenem in the primary efficacy analyses (5). However, the subgroup of patients with moderate renal impairment at baseline (CL_CR of 31 to 50 ml/min) had lower clinical cure rates in both groups than the overall population, and there was a disproportionate decrease in cure rates for ceftazidime-avibactam compared with meropenem (5). A potential period of underdosing in a proportion of renally impaired patients who had rapidly improving renal function during the early treatment period was thought to have contributed to this finding (5). Of note, this was not replicated across the other ceftazidime-avibactam phase 3 trials.

Ceftazidime-avibactam dosages for initial phase 3 investigation did not account for rapidly improving renal function and were selected based on relatively limited population PK data, as well as existing label information for ceftazidime. Following the findings from RECLAIM, modified ceftazidime-avibactam dosage regimens for patients with CL_CR of <50 ml/min, which aimed to reduce the risk of underdosing in the event of rapid improvements in renal status, were proposed based on updated exposure and PTA analyses. The modified dosage regimens were predicted to maintain exposures comparable to those in patients with normal renal function or mild renal impairment receiving the standard dosage regimen and to achieve >90% PTA across indications and renal function groups. Although it is not possible to select dosage regimens that would provide >90% PTA in the event of rapidly improving renal function without overexposing patients with stable renal function, the modified dosages increased PTAs in the event of renal function improvement to the immediately higher renal function category (without dosage change) to 66.4 to 97.0% (16). In addition, >50% PTA was predicted for the modified dosage regimen in the event of rapid improvement in renal function from moderate impairment (CL_CR of 30 to 50 ml/min) to normal renal function (CL_CR of >80 ml/min) (16).

As the remaining phase 3 trials were ongoing at the time the RECLAIM findings were made known, the protocols of some ongoing trials were amended to restrict the enrollment of patients with CL_CR of <50 ml/min during the evaluation and implementation of the revised dosages, and the modified dosages were ultimately implemented partway through the final phase 3 REPROVE trial in patients with NP/VAP. However, reduced efficacy for ceftazidime-avibactam versus comparators was not observed for patients with CL_CR of ≤50 ml/min receiving the original dosage regimens in the other trials (6–9). The physiological derangements (for example, hypovolemia and hemodynamic instability) and fluid shifts that occur in the setting of severe infection and surgery in patients with cIAI can result in decreased renal perfusion at initial presentation (reflected by increased SCr and decreased estimated CL_CR), which may rapidly improve in response to fluid resuscitation and early intervention in the treatment of infection. The outcomes for moderate renal impairment in RECLAIM might therefore be specific to the cIAI indication, although similar findings were not seen in RECLAIM 3 (8), possibly because of the lower numbers of patients with baseline CL_CR of <50 ml/min. The REPROVE study protocol was amended to reflect the modified dosage regimens described here, but 62 patients with CL_CR of <50 ml/min were enrolled before the protocol amendment (9). Clinical cure rates and 28-day all-cause mortality for patients in REPROVE with baseline CL_CR of ≤50 ml/min receiving either the original or the modified dosage regimens were similar to those in the overall populations (9, 24, 25), thus supporting the hypothesis that potential underdosing for the original dosage adjustments was specific to the cIAI indication.

The completion of the ceftazidime-avibactam phase 3 program provided the opportunity to further update the population PK models and validate the previous exposure and PTA estimates (17). These analyses confirmed the previously predicted high PTAs (>90%) for the modified dosage adjustments across indications. Individual-predicted exposures for patients in the phase 3 studies using the final models confirmed the expected exposures predicted by the earlier models and indicated that PK/PD target attainment >95% was maintained in patients with potential ARC, such that no ceftazidime-avibactam dosage adjustments are required in these patients (17).

The modified dosages for renal impairment described in this review have been incorporated within U.S. and European product labeling for ceftazidime-avibactam (1, 2). In patients with impaired renal function, regular monitoring of estimated CL_CR is advised, as in some patients the CL_CR estimated from SCr can change quickly, especially early in the course of treatment (1, 2). Ceftazidime-avibactam is a relatively new therapeutic option for serious Gram-negative infections, and given the potential adverse outcomes of treatment failure in this setting, there is understandable interest in ensuring that dosages are optimal for all patient groups, including those with renal impairment (31, 32, 35). The results summarized here give reassurance that the approved dosage regimens provide adequate exposure and PTA for patients with a range of renal function from severe impairment/ESRD to ARC while remaining within established safety exposure limits. This experience with the ceftazidime-avibactam clinical development program, regarding the value of population PK modeling to proactively evaluate and modify dosage regimens as clinical data accumulate, has broader implications for all therapeutic drugs that are cleared by the kidneys and for other antibiotics for serious infections, for which rapid changes in renal function during treatment may be encountered. Moreover, had the ceftazidime-avibactam combination not been investigated in patients with cIAI, the issue of potential ceftazidime-avibactam underexposure in patients with rapidly improving renal function in the RECLAIM program would not have been identified and successfully addressed with the modified dosage adjustments that have been carried forward into the current label information.

CONCLUSIONS

Ceftazidime-avibactam dosage selection, modification, and validation have been performed iteratively during the clinical development process using an extensive and robust population PK modeling approach and PK/PD targets to ensure that recommended dosing regimens for all categories of renal function provide adequate exposures and high predicted PTA values. The combination has demonstrated consistent clinical efficacy equivalent to those of carbapenem-based comparators in five adult phase 3 trials across multiple indications (5–9). Modified ceftazidime-avibactam dosage adjustments for CL_CR of ≤50 ml/min, wherein the total daily dose was increased within established safety exposure limits for patients with normal renal function (CL_CR of >80 ml/min), were developed during the phase 3 program to mitigate the risk of rapidly improving renal function causing suboptimal exposures. These dosing adjustments have been incorporated within the current European and U.S. FDA dosing recommendations for ceftazidime-avibactam (1, 2).