A Review of Extended and Continuous Infusion Beta-Lactams in Pediatric Patients

Abstract

Intravenous beta-lactam antibiotics are the most prescribed antibiotic class in US hospitalized patients of all ages; therefore, optimizing their dosing is crucial. Bactericidal killing is best predicted by the time in which beta-lactam drug concentrations are maintained above the organism's minimum inhibitory concentration (MIC), rather than achievement of a high peak concentration. As such, administration of beta-lactam antibiotics via extended or continuous infusions over a minimum of 3 hours, rather than standard infusions over approximately 30 minutes, has been associated with improved achievement of pharmacodynamic targets and improved clinical outcomes in adult medical literature. This review summarizes the pediatric medical literature. Applicable studies include pharmacodynamic models, case series, retrospective analyses, and prospective studies on the use of extended infusion and continuous infusion penicillins, cephalosporins, carbapenems, and monobactams in neonates, infants, children, and adolescents. Specialized patient populations with unique pharmacokinetics and high-risk infections (neonates, critically ill, febrile neutropenia, cystic fibrosis) are also reviewed. While more studies are needed to confirm prospective clinical outcomes, the current body of evidence suggests extended and continuous infusions of beta-lactam antibiotics are well tolerated in children and improve achievement of pharmacokineticpharmacodynamic targets with similar or superior clinical outcomes, particularly in infections associated with high MICs.

Introduction

Parenteral beta-lactams, including penicillins, cephalosporins, carbapenems, and monobactams, are the most commonly prescribed antibiotics for empiric and definitive treatment of infections in hospitalized US adults and children.^1,2 Beta-lactams demonstrate broad yet distinct spectrums of antimicrobial activity directed against Gram-positive, Gram-negative, and/or anaerobic pathogens. Pharmacokinetic and pharmacodynamic (PKPD) research have established that beta-lactams exhibit optimal bacterial killing and microbiologic response as a function of time in which the free beta-lactam concentration is maintained above the organism's minimum inhibitory concentration (MIC), or fT > MIC.^3–5 Maintaining antibiotic concentrations above the organism's MIC in vivo may be challenging, particularly for patients with altered PK, organisms with elevated MICs, infections in difficult-to-penetrate locations, or antibiotics with short elimination half-lives.⁶

Children have specific PK-PD parameters, including expanded volume of distribution for hydrophilic medications and rapid renal clearance, which decrease the likelihood of achieving beta-lactam PK-PD targets.^6–8 Thus, this population is a prime target for augmented dosing strategies. Three chief dosing strategies have been used in an effort to maximize target attainment: 1) administering the antibiotic more frequently; 2) prolonging the infusion time from a standard infusion (SI)—typically over 30 to 60 minutes—to an extended infusion (EI)—typically over 3 to 4 hours; or 3) infusing the antibiotic over 24 hours as continuous infusion (CI).

In adult populations, PD modeling and patient-level data support EI dose augmentation,^9–13 prompting many institutions to administer beta-lactam antibiotics via EI as standard of care.^14–16 Published benefits of EI or CI in adults include faster time to defervescence, enhanced cost effectiveness, shorter intensive care unit (ICU) length of stay, higher rate of clinical cure, and decreased mortality.^{9–11,17–21} This review aims to summarize the available literature for the use of EI and CI beta-lactams in pediatric patients.

Pharmacokinetics and Pharmacodynamics of Beta-Lactam Antibiotics

To apply EI and CI data to the pediatric population, it is important to understand the PK-PD parameters of beta-lactams. While it is established that the duration of time in which the free (non–protein bound) beta-lactam concentration exceeds the organism's MIC best predicts bacterial killing and microbiologic response, the specifics of this goal, particularly how high above the MIC and for what percentage of the dosing interval, remain active areas of discussion.^22–25 Maintaining beta-lactam serum concentrations above the MIC for approximately 50% of the dosing interval (40% for carbapenems, 50%–60% for penicillins/monobactams, 60% for cephalosporins) results in bactericidal activity in vitro.^24,26–28 These targets represent the traditional, long-established goal parameters for beta-lactam therapeutic drug monitoring (TDM), and achievement of these targets in vivo has been associated with improved survival^29,30 and microbiologic success³¹ in adult patients.

A more aggressive PD target—100% fT > MIC—has also been correlated with clinical success. This target was associated with higher rate of positive outcomes (completion of treatment without change or addition of antibiotic), compared with 50% fT > MIC, in a prospective study of critically ill adults.³² While this analysis has several limitations including lack of standard regimens, randomization, and sometimes an identified bacterial pathogen, this study provides real-world correlation of response with achievement of prespecified PD parameters. Similarly, achievement of cefepime or ceftazidime 100% fT > MIC in septic adults was linked to statistically significant increased rates of clinical cure and increased bacterial eradication vs <100% fT > MIC.^33,34

Lastly, some authors argue that serum beta-lactam minimum concentration should be sustained at 4 to 6 times the MIC to optimize both clinical and bacteriologic response.^22,35,36 An investigation of critically ill children undergoing extracorporeal membrane oxygenation (ECMO) or continuous renal replacement therapy (CRRT) demonstrated that 95% of children did not achieve 40% fT > 4–6× MIC with standard dosing. With dose adjustments to achieve this target, ultimately 22 of 23 (95.7%) patients with confirmed infections clinically improved.³⁷ Similarly, 3 of 4 critically ill children undergoing continuous veno-venous hemodiafiltration and receiving cefepime achieved 100% fT > MIC, while only 1 patient achieved 100% fT > 4× MIC, demonstrating the challenge associated with achieving high PD targets and the potential benefits of beta-lactam TDM.³⁸ This higher PK-PD target has also been associated with preventing selection of resistant organisms.^{33,34,39–42} These studies must be interpreted with their limitations in mind, because more prospective clinical data for optimal PK-PD target are needed, particularly in children.

Monte Carlo simulations are mathematical constructs using computer software to discern how modifying elements of an antimicrobial drug regimen, including dose, frequency, or infusion time, affects the achievement of PD targets for a given patient population.^43,44 These models include patient-specific variables such as age, renal function, acuity of illness, infection type, or comorbidities and often simulate large sample sizes— making them especially useful in researching novel dosing regimens and patient populations with limited representation in clinical studies. As such, Monte Carlo simulations have been used to evaluate EI beta-lactam regimens and are described throughout this review.^45,46

Penicillins

Piperacillin-tazobactam. Piperacillin has a half-life of approximately 42 minutes in children,⁴⁷ which is notably shorter than that of other antipseudomonal beta-lactams (i.e., cefepime 1.8 hours,⁴⁸ meropenem 1 hour,49 aztreonam 1.7 hours).⁵⁰ This rapid clearance results in lower likelihood of sustained drug concentrations over time when using infrequent or short infusions. Across 5 Monte Carlo studies totaling 37,000 simulated children, 1 studied regimen unanimously achieved goal PK-PD targets—piperacillin-tazobactam 100 to 120 mg/kg every 6 hours EI.^46,51–53 Other EI regimens also frequently achieved goal, including 80 mg/kg every 6 hours EI (2 of 2 simulations) and 100 to 130 mg/kg every 8 hours EI (3 of 4 simulations). All the CI regimens, ranging from piperacillin-tazobactam 300 to 480 mg/kg/day CI, achieved PK-PD target as well. On the other hand, SI regimens were less likely to achieve target parameters. No SI regimens dosed every 8 hours achieved PK-PD target across these 5 analyses. Standard infusions administered every 6 hours were largely unsuccessful as well (2 of 7 patient groups across studies achieved goal), and every 4 hours SIs resulted in mixed success. These highlighted studies all used the same, most traditional, PK-PD target of ≥50% fT > MIC and the standard Clinical and Laboratory Standards Institute (CLSI) susceptible breakpoint of 16 mg/L, with patients between the ages of 1 to 12 years.^46,51–53 Therefore, these studies are largely applicable to pediatric clinical practice.

More stringent PK-PD targets, such as ≥100% fT > MIC, were only achieved with use of CI doses ≥300 mg/kg/day in 2 other Monte Carlo models.^54,55 These data illustrate the value of augmented regimens and the interplay between concentration, drug PD goals, and organism MIC.

Observational and prospective clinical studies have also evaluated piperacillin-tazobactam EI outcomes in children. Knoderer et al⁵⁶ conducted a retrospective case series of 50 children, ages 2 to 9 years, receiving a piperacillin-tazobactam EI (Table 1). All children had Gram-negative organism(s) in cultures of the blood, bronchoscopy fluid, urine, wound, or more than 1 of these sites. The patient cohort achieved 100% microbiologic cure and 74% clinical cure at 21 days. No patients experienced adverse effects related to the regimen. While this study is limited by the lack of a SI comparator group, it demonstrates feasibility of piperacillin-tazobactam EI as standard of care in a pediatric hospital, safety/tolerability of EI, and generally favorable clinical outcomes.⁵⁶ In addition, a randomized, controlled, non-blinded trial compared piperacillin-tazobactam SI with CI for the treatment of febrile neutropenia.⁵⁷ Seventy-six children received the CI regimen. No statistically significant difference in treatment failure was detected between groups: 13% vs 21% (p = 0.15). The low rate of proven bacteremia (8%) and the high rate of fever of unknown origin diagnoses may have contributed to these findings. No adverse effects are noted in the publication; however, 4 patients were excluded from CI owing to inadequate intravenous (IV) access.⁵⁷ Future prospective studies of pediatric patients with confirmed Gram-negative infections may be better designed to detect differences in clinical outcomes. Nevertheless, available data indicate that EI piperacillin-tazobactam increases the likelihood of achieving PK-PD targets in infants and children with minimal risks of adverse events.

Table 1.

Extended and Continuous Infusions of Penicillins in Pediatrics

Reference (Population)	PK-PD Goal	Regimens Studied*	MIC, mg/L	PTA (Outcome)
Piperacillin-tazobactam (doses reported in mg/kg piperacillin)
Ellis45 (5000‡; 10 yr males; GP; ICU)	≥50% fT > MIC	75 mg/kg every 6 hr SI (max: 4000 mg/dose)	4†	60% for Institution 1
				74% for Institutions 2
Courter46 (5000‡; 2 and 12 yr males; GP)	≥50% fT > MIC	80 mg/kg every 6 hr SI, EI3	16	48% SI, 100% EI
		100 mg/kg every 6 hr SI, EI3		58% SI, 100% EI
		120 mg/kg every 6 hr SI, EI3		66% SI, 100% EI
		100 mg/kg every 8 hr SI, EI3		21% SI, 83% EI
		300–480 mg/kg CI		100% (all doses)
Cies51 (5000‡; 13§; 1–6 yr males; ICU)	≥50% fT > MIC	50 mg/kg every 4 hr SI	16	<90% SI
		100 mg/kg every 6 hr SI, EI3		<90% SI, >90% EI
		80 mg/kg every 8 hr SI, EI4		<90% SI, <90% EI
		400 mg/kg CI		>90%
Nichols52 (12,000‡; 12§; 2–8 yr; ICU)	≥50% fT > MIC	80 mg/kg every 6 hr SI, EI3	16	95% SI, 100% EI
		100 mg/kg every 6 hr SI, EI3		98% SI, 100% EI
		80 mg/kg every 8 hr SI, EI4		60% SI, 100% EI
		100 mg/kg every 8 hr SI, EI4		75% SI, 100% EI
	100% fT > MIC	80 mg/kg every 6 hr SI, EI3		10% SI, 45% EI
		100 mg/kg every 6 hr SI, EI3		0% SI, 60% EI
		80 mg/kg every 8 hr SI, EI4		10% SI, 10% EI
		100 mg/kg every 8 hr SI, EI4		0% SI, 15% EI
Cies53 (5000‡; 21§; 3–10 yr; FN)	≥50% fT > MIC	50 mg/kg every 4 hr SI	16	94%
		100 mg/kg every 6 hr SI, EI3 (max: 4000 mg/dose)		87% SI, 100% EI
		100 mg/kg every 8 hr SI, EI4		65% SI, 100% EI
		400 mg/kg CI		100%
Delvallée54 (N = 19; 7–14.5 yr; FN)	100% fT > 6 × MIC	300 mg/kg CI preceded by LD¶	4	82%–87%
		350 mg/kg CI preceded by LD¶		100%
		400 mg/kg CI preceded by LD¶		100%
Maarbjerg55 (N = 100,000‡; 43§; 1–18 yr; FN)	≥100% fT > MIC	75 mg/kg every 6 hr SI, EI3	16	0% SI, 0% EI
		100 mg/kg every 6 hr SI, EI3		0% SI, 1.1% EI
		100 mg/kg every 8 hr SI, EI4		0% SI, 0% EI
		133.3 mg/kg every 8 hr SI, EI4		0% SI, 0% EI
		300, 400 mg/kg CI (max: 16,000 mg)		100% (both doses)
	50% fT > 4 × MIC	75 mg/kg every 6 hr SI, EI3	8	<95% SI, <95% EI
		100 mg/kg every 6 hr SI, EI3		<95% SI, >95% EI
		100 mg/kg every 8 hr SI, EI4		<95% SI, <95% EI
		133.3 mg/kg every 8 hr SI, EI4		<95% SI, <95% EI
		300, 400 mg/kg CI (max: 16,000 mg)		>95% (both doses)
Thibault95 (N = 1000‡; 15§; 2–5 mo; FN; GP; ICU)	≥50% fT > MIC	75 mg/kg every 4 hr SI	16	>90%
		80 mg/kg every 6 hr EI2		<90%
		120 mg/kg every 6 hr EI3		>90%
		80 mg/kg every 8 hr EI4		>90%
		130 mg/kg every 8 hr EI4		>90%
		300 mg/kg CI		>90%

Table 1.

Extended and Continuous Infusions of Penicillins in Pediatrics (cont.)

Reference (Population)	PK-PD Goal	Regimens Studied*	MIC, mg/L	PTA (Outcome)
Thibault95 (N = 1000‡; 15§; 2–5 mo; FN; GP; ICU)	≥50% fT > MIC	75 mg/kg every 4 hr SI	16	>90%
		80 mg/kg every 6 hr EI2		<90%
		120 mg/kg every 6 hr EI3		>90%
		80 mg/kg every 8 hr EI4		>90%
		130 mg/kg every 8 hr EI4		>90%
		300 mg/kg CI		>90%
Thibault95 (N = 1000‡; 74§; 6 mo–6 yr; FN; GP; ICU)	≥50% fT > MIC	75 mg/kg every 4 hr SI	16	<75%
		120 mg/kg every 6 hr EI3		>90%
		80 mg/kg every 8 hr EI4		<75%
		90 mg/kg every 8 hr EI4		<75%
		130 mg/kg every 8 hr EI4		>90%
		400 mg/kg CI		>90%
Knoderer56 (N = 50; 2–9 yr; FN; GP; ICU)	N/A	89–100 mg/kg every 8 hr EI4	N/A	(74% clinical and 100% microbiologic cure)
Solórzano-Santos57 (N = 176; 1–16 yr; FN)	N/A	50 mg/kg every 6 hr SI	N/A	(13% Tx failure)
		300 mg/kg CI preceded by		(21% Tx failure)
		75 mg/kg LD
Zembles66 (N = 551#; 3–14 yr ; FN, GP; ICU)	N/A	100 mg/kg every 8 hr SI, EI4** (max: 4000 mg/dose)	N/A	(No statistical difference in LOS, readmission, mortality. Lower mortality in ICU (p = 0.006) and FN readmission with EI (p = 0.012))
Nafcillin
Prazak61 (N = 4; 4–8 yr; Pneumonia)	≥100% fT > MIC	200 mg/kg CI	≤0.5	100%
Knoderer62 (N = 40; 2.3–12 yr; GP)	N/A	190 ± 36.4 mg/kg CI	N/A	(92% Tx success; 91% microbiologic cure; 20% rate of ADR

ADR, adverse drug reactions; CI, continuous infusion; EI, extended infusion; EI₃, extended infusion over 3 hrs; EI₄, extended infusion over 4 hrs; FN, febrile neutropenia; GP, general pediatrics; ICU, intensive care unit; LD, loading dose; max, maximum; LOS, length of stay; MIC, minimum inhibitory concentration; N/A, not applicable/available; PK-PD, pharmacokinetic-pharmacodynamic; PTA, probability of target attainment; SI, standard infusion; Tx, treatment

* Bolded regimens are the preferred regimens per study authors.

^† Median MIC reported.

^‡ Number of simulations.

§ Number of unique patients who contributed serum concentrations.

¶ Simulated loading doses included 50 mg/kg, 100 mg/kg, and 150 mg/kg for each CI regimen.

# Total number of patients who were retrospectively included in the study if they received standard or prolonged infusion piperacillin-tazobactam, cefepime, or meropenem for at least 72 hours.

** Clinical outcomes for individual antibiotic agents (piperacillin-tazobactam, cefepime, or meropenem) were not reported

Nafcillin. Similarly to piperacillin-tazobactam, nafcillin's half-life is exceptionally short (adults 0.5–1 hour; children 0.5–1.9 hours), resulting in frequent (every 4 hours) intermittent dosing to maintain time-dependent killing.^58,59 While nafcillin CI literature is limited in adults, its use is supported by the Infectious Diseases Society of America for the treatment of vertebral osteomyelitis.⁶⁰ In children, Prazak et al⁶¹ assessed PK parameters of 4 burn patients treated with nafcillin CI for methicillin-susceptible Staphylococcus aureus (MSSA) pneumonia (Table 1). Their findings demonstrate the feasibility of CI, and all 4 patients achieved PK-PD target.⁶¹ A larger series of 40 pediatric patients receiving nafcillin CI exhibited favorable clinical responses with prescriber-documented treatment success and microbiologic cure achieved in over 90% of patients.⁶² The regimen was well tolerated with 20% of patients experiencing any adverse event (no SI comparator group).⁶² Further studies directly comparing nafcillin SI with CI are needed, but existing evidence indicates nafcillin CI reliably achieves PK-PD targets and may offer practical advantages.

Cephalosporins

Cefepime. Across 3 cefepime Monte Carlo simulations for infants and children, the only 2 regimens to achieve the traditional PK-PD goal of 50% to 60% fT > MIC were EI and CI regimens.^45,46,63 Even when administered every 6 hours, SI cefepime failed to meet this conservative target, although only by a small margin (Table 2). Notably, the cefepime pediatric dose per manufacturer prescribing information is 50 mg/kg IV every 8 or 12 hours infused over 30 minutes, which has been associated with probability of target attainment (PTA) as low as 64% and 15%, respectively.^45,46,63,64

Table 2.

Extended and Continuous Infusions of Cephalosporins in Pediatrics

Reference (Population)	PK-PD Goal	Regimens Studied*	MIC, mg/L	PTA (Outcome)
Cefepime
Ellis45 (N = 5000‡; 10 yr males; GP; ICU)	≥50% fT > MIC	50 mg/kg every 8 hr SI (max: 2000 mg/dose)	4†	85% for institution 1 78% for institution 2
Courter46 (N = 5000‡; 2 and 12 yr males; GP)	≥50% fT > MIC	50 mg/kg every 8 hr SI, EI3	8	79% SI, 100% EI
		50 mg/kg every 12 hr SI, EI3		15% SI, 57% EI
		100, 150 mg/kg CI		100%
Shoji63 (N = 44§; <30 days; NICU)	≥60% fT > MIC	30 mg/kg every 12 hr SI	8	>90% preterm;
		50 mg/kg every 12 hr SI		<90% term infants
				>90% preterm and term infant
Shoji63 (N = 47§; 1 mo–11 yr; GP)	≥60% fT > MIC	50 mg/kg every 6 hr SI	8	89%
		50 mg/kg every 8 hr SI, EI3		69% SI, 92% EI
		50 mg/kg every 12 hr SI		32%
Nichols65 (N = 150; 2–12.3 yr; FN, GP, ICU)	N/A	50 mg/kg every 8 hr EI4 (max: 1000 mg/dose for most indications; 2000 mg/dose for CNS infections or weight >100 kg)	N/A	93% remained on EI (8% medication incidents reported)
Zembles66 (N = 551¶; 3–14 yr ; FN, GP, ICU)	N/A	50 mg/kg every 8 hr SI, EI4# (max: 2000 mg/dose)	N/A	(No significant difference in LOS, readmission, mortality. Lower mortality in ICU (0.006) and lower readmission in FN (p = 0.012) with EI)
Ceftazidime
Ellis45 (N = 5000‡; 10 yr males; GP, ICU)	≥50% fT > MIC	50 mg/kg every 8 hr SI (max: 2000 mg/dose)	2†	92% for institution 1
			4†	65% for institution 2
Courter46 (N = 5000‡; 2 and12 yr males; GP)	≥50% fT > MIC	30 mg/kg every 8 hr SI, EI3	8	54% SI, 93% EI
		50 mg/kg every 8 hr SI, EI3		80% SI, 100% EI
		100, 150 mg/kg CI		100% (both doses)
Cojutti68 (N=1000‡; 46§; 6 mo–16 yr; FN)	Css ≥4× MIC	60–100 mg/kg LD followed by	8	94% with 200 mg/kg CI; >90% with
		100–200 mg/kg (or 1–6 g)		4–6 g CI based upon BSA and eGFR
Dalle70 (N = 20; 1–15 yr; FN)	Css ≥ 5× MIC	65 mg/kg LD (max: 2 g) followed by 200 mg/kg CI (max: 6 g) plus amikacin and vancomycin	N/A	100% for mean values at all time points. (Patients tolerated and preferred CI. No toxic or infectious mortality.)
David69 (N = 9; Age N/A; CF)	N/A	100 mg/kg LD followed by 300 mg/kg CI	N/A	100% fT >13 × MIC
Rappaz71 (N = 14; 2 consecutive exacerbations; 5–16.8 yr; CF)	≥100% fT > MIC	66.7 mg/kg every 8 hr SI plus amikacin	0.5–4	68% SI, 100% CI. (All patients improved clinically. No difference in changes in FEV1. Prealbumin increased more in CI (p = 0.015). Patients tolerated and preferred CI.)
		100 mg/kg CI plus amikacin	N/A	(All concentrations ≥ MIC and all patients improved clinically)

Table 2.

Extended and Continuous Infusions of Cephalosporins in Pediatrics (cont.)

Reference (Population)	PK-PD Goal	Regimens Studied*	MIC, mg/L	PTA (Outcome)
Riethmueller72 (N = 56; 5–37 yr; CF)	N/A	66.7 mg/kg every 8 hr SI plus tobramycin	N/A	(No significant differences in FEV1, body weight, or PsA density.)
		100 mg/kg CI plus tobramycin
Hubert73 (N = 70; 16–31 yr; CF)	N/A	66.7 mg/kg every 8 hr SI 60 mg/kg LD (max: 2 g) followed by 200 mg/kg CI (max: 12 g)	N/A	(No significant difference in change in FEV1 between groups. Significantly > change in FEV1 if MIC ≥32 mg/mL with CI (p < 0.05). Longer time between exacerbations with CI (p = 0.04). No difference in ADR. Similar quality-of-life scores 47 of 57 preferred CI for future treatment.)

ADR, adverse drug reactions; BSA, body surface area; CF, cystic fibrosis; CI, continuous infusion; CNS, central nervous system; C_ss, steady-state concentration; eGFR, estimated glomerular filtration rate; EI, extended infusion; EI₃, extended infusion over 3 hrs; EI₄, extended infusion over 4 hrs; FEV1, forced expiratory volume in 1 second; FN, febrile neutropenia; GP, general pediatrics; ICU, intensive care unit; LD, loading dose; LOS, length of stay; max, maximum; MIC, minimum inhibitory concentration; N/A, not applicable/available; NICU, neonatal ICU; PK-PD, pharmacokinetic-pharmacodynamic; PsA, Pseudomonas aeruginosa; PTA, probability of target attainment; SI, standard infusion

* Bolded regimens are the preferred regimens per study authors.

^† Median MIC reported.

^‡ Number of simulations.

§ Number of unique patients who contributed serum concentrations.

¶ Total number of patients who were retrospectively included in the study if they received standard or prolonged infusion piperacillin-tazobactam, cefepime, or meropenem for at least 72 hours.

# Clinical outcomes for individual antibiotic agents (piperacillin-tazobactam, cefepime, or meropenem) were not reported.

Importantly, one study simulated cefepime PK-PD in neonates, including premature infants with gestational age (GA) <36 weeks. This young population more easily achieved PK-PD goals on SI regimens such that cefepime 50 mg/kg every 12 hours SI resulted in 99% PTA in GA <36 weeks and 95% PTA in GA ≥36 weeks.⁶³ Therefore, neonates are less likely to benefit from cefepime EI than older children.

The first feasibility study describing cefepime EI use in pediatrics was published by Nichols et al⁶⁵ in 2015. Cefepime 50 mg/kg IV every 8 hours EI (max 1 g/dose for patients <100 kg without central nervous system infection) was prospectively administered to children ages 1 month to 17 years. One hundred fifty acutely and critically ill children received cefepime EI, with 93% of patients completing treatment course without regimen change.⁶⁵ The most common reasons for resuming SIs were drug incompatibility or limited IV access.⁶⁵ Recently, Zembles et al⁶⁶ reported clinical outcomes of 551 children who received cefepime, piperacillin-tazobactam, or meropenem by either SI or EI and found no differences in 30-day all-cause mortality, median length of stay, or readmission. However, a priori subanalysis revealed decreased mortality in critical care patients (2.1% vs 19.6%, p = 0.006) and decreased readmissions in the hematopoietic stem cell transplant (HSCT) population (0% vs 50%, p = 0.012) who received EI vs SI.

Ceftazidime. Ceftazidime's conventional PK-PD goal of 50% fT > MIC has been used to develop 2 pediatric Monte Carlo simulations. One study observed PK-PD target attainment using isolates from 2 different institutions; the hospital with a lower median organism MIC achieved >90% PTA with SI dosing, while the hospital with a higher median organism MIC did not (Table 2).⁴⁵ Importantly, both median MICs were below the CLSI clinical breakpoint of ≤8 mg/L for Pseudomonas aeruginosa; therefore, it is important to consider if time-dependent killing can be achieved—even when interpretations are reported as “susceptible.”⁶⁷ In the second PK-PD model, only EI and CI regimens achieved target (Table 2).⁴⁶ A third simulation modeled ceftazi-dime CI in pediatric HSCT recipients to target steady-state concentration ( C_ss ) ≥ 4× MIC, for which authors advise doses of 200 mg/kg/day or 4 to 6 g/day CI.^68,69

Patient PK-PD target attainment, feasibility, and tolerability were evaluated in children receiving ceftazidime CI for febrile neutropenia (FN). Twenty-three episodes of FN were treated for a median duration of 7 days (range, 7–21 days), during which no infectious bacteria were isolated.⁷⁰ All mean C_ss achieved PK-PD target with no instances of toxic or infectious mortality (Table 2). Treatment was well tolerated by all children, and authors note that CI was simpler for caretakers and more comfortable for patients.⁷⁰

Clinical outcomes using augmented dosing strategies have also been evaluated in the pediatric cystic fibrosis (CF) population. Rappaz et al⁷¹ sequentially treated 14 children (mean 12.6 years) with chronic P aeruginosa infection with SI during a pulmonary exacerbation, followed by CI during a subsequent exacerbation, each for 14 days' duration and in combination with an aminoglycoside. Serum ceftazidime concentrations were consistently maintained above the organism MIC with CI vs only two-thirds of samples with SI. While clinical response and change in forced expiratory volume in 1 second (FEV1) were equivalent between dosing strategies, patients favored ceftazidime CI.⁷¹ The same dosing regimens were also compared in a randomized crossover study including 56 CF patients (mean age 14.4 years); however, patients received routine, elective courses of antibiotics in the context of chronic infection rather than for acute pulmonary exacerbation.⁷² Significant improvements in FEV1, lung bacterial burden, weight, and leukocyte count were achieved by both SI and CI treatment groups.⁷² Finally, a randomized multicenter crossover study treated patients ages 8 years and older with ceftazidime SI or CI for 2 consecutive pulmonary exacerbations. Change in FEV1 was not statistically different between groups (+ 5.5% for SI and +7.6% for CI, p = 0.15).⁷³ However, amongst patients harboring P aeruginosa with MIC >32 mg/L, change in FEV1 was significantly different between groups (+1.7% for SI and +6.2% for CI, p < 0.05), favoring CI. Time between pulmonary exacerbation treatment courses was longer for patients after receiving CI (SI 2.8 months; CI 3.2 months; p = 0.04). At study completion, 82% of patients preferred ceftazidime CI for future treatment courses.⁷³ Collectively, these studies suggest that ceftazidime CI is associated with similar clinical efficacy, comparable tolerability, enhanced PTA, greater patient satisfaction, and potential for prolonging time to next exacerbation when compared with SI dosing in CF.

Carbapenems

Meropenem. Monte Carlo simulations, case reports, and prospective studies have described meropenem EI use in various pediatric age groups, disease states, and infection types. Conventional PK-PD goal of 40% fT > MIC was used in 2 pediatric models, which demonstrated that meropenem EI consistently achieved target, while SI achieved target in one institution with a low median organism MIC, but not in another institution (Table 3).^45,46 Using a higher target of 70% fT > MIC, Wang and colleagues⁷⁴ evaluated serum concentrations from 57 pediatric ICU patients receiving meropenem for bacterial meningitis, sepsis, or severe pneumonia and developed a Monte Carlo model. In this model, meropenem SI failed to achieve 90% PTA; EI regimens, while closer to the target parameter, also did not meet 90% PTA. Continuous infusion successfully achieved target PTA, even at the highest MICs tested (Table 3).⁷⁴

Table 3.

Extended and Continuous Infusions of Carbapenems in Pediatrics

Reference (Population)	PK-PD Goal	Regimens Studied*	MIC, mg/L	PTA (Outcome)
Meropenem
Ellis45 (N = 5000‡; 10 yr males; GP, ICU)	≥40% fT > MIC	20 mg/kg every 8 hr SI (max: 2000 mg/dose)	1†	84% for institution 1
			4†	47% for institution 2
		40 mg/kg every 8 hr SI (max: 2000 mg/dose)	1†	92% for institution 1
			4†	58% for institution 2
Courter46 (N = 5000†; 2 yr, 12 yr; GP)	≥40% fT > MIC	20 mg/kg every 8 hr SI, EI3	4	33% SI, 97% EI
		40 mg/kg every 8 hr SI, EI3		72% SI, 100% EI
Wang74 (N = 1000†; 57§; 1 mo–14 yr; ICU)	≥70% fT > MIC	20 mg/kg every 8 hr SI, EI4	1	19% SI, 70% EI
		20 mg/kg every 8 hr SI, EI4	2	6% SI, 40% EI
		40 mg/kg every 8 hr EI4	1	87%
		40 mg/kg every 8 hr EI4	2	69%
		110 mg/kg CI	4	98%
		110 mg/kg CI	8	73%
Ohata77 (N = 1000†; 154§; 2–3.5 yr; Meningitis)	CSF ≥50% fT > MIC	20 mg/kg every 8 hr SI, EI4	0.12	73% SI, 75% EI
		40 mg/kg every 8 hr SI, EI4		86%, SI 90% EI
Cojutti82 (N = 10,000†; 21§; 4.2–15 yr; HSCT)	100% fT > 4× MIC	45–60 mg/kg CI preceded by LD	2	≥90% for CrCL120–299**
		90 mg/kg CI preceded by LD	8	≥90% for CrCL 40–119**
Avedissian98 (N = 1,000†; 105§; 4.4–15.2 yr; ICU)	40% fT > 4× MIC	20 mg/kg every 8 hr SI	2	>90% for CrCL 100–160**
		40 mg/kg every 8 hr SI		>90% CrCL 100–160**
	80% fT > 4× MIC	20 mg/kg every 8 hr SI		<90% for CrCL 100; <50% for CrCL
		40 mg/kg every 8 hr SI		140; and <30% for CrCL 160**
				>90% for CrCL 100; <80% for CrCL
				140; 60% for CrCL 160**
Padari75 (N = 19; GA ≤ 32 wk; NICU)	≥100% fT > MIC	20 mg/kg every 12 hr SI, EI4	2	100% SI, 99.9% EI
	fT > 6× MIC	20 mg/kg every 12 hr SI, EI4		80% SI, 82% EI
Shabaan76 (N = 102); Neonates; NICU)	N/A	20 mg/kg SI, EI4	N/A	(Significant differences in clinical improvement (p = 0.009), microbiologic eradiation (p = 0.009), mortality (p = 0.03), duration of respiratory support (p = 0.03), and acute kidney injury (p = 0.02) favoring EI group.)
		40 mg/kg SI, EI4 for meningitis or PsA
Pettit79 (N = 5000†; 30§; 6–17 yr; CF)	≥40% fT > MIC	40 mg/kg every 8 hr SI, EI3 (max: 2000 mg/dose)	1	88% SI, >99% EI (Well tolerated; 18 of 30 completed entire course as EI [13 hospital, 5 outpatient])
		40 mg/kg every 8 hr SI, EI3	2	70% SI, >99% EI
		40 mg/kg every 8 hr SI, EI3	4	35% SI, >99% EI
		40 mg/kg every 8 hr SI, EI3	8	83% EI
Kuti80 (N = 15; 8–17 yr; CF)	≥40% fT > MIC	40 mg/kg every 8 hr EI3 (max: 2000 mg/dose)	0.125–16 for PsA	(63% fT > MIC was associated with improvement in FEV1 [p < 0.0004])
Zembles66 (N = 551¶; 3–14 yr; FN; GP; ICU)	N/A	20–40 mg/kg every 8 hr SI, EI3# (max: 2000 mg/dose)	N/A	(No significant difference in overall clinical outcomes: LOS, readmission, mortality. Lower mortality in ICU (p = 0.006) and readmission in FN (p = 0.012) with EI)

Table 3.

Extended and Continuous Infusions of Carbapenems in Pediatrics (cont.)

Reference (Population)	PK-PD Goal	Regimens Studied*	MIC, mg/L	PTA (Outcome)
Imipenem/cilastatin
Ellis45 (N = 5000†; 10 yr males GP, ICU)	≥40% fT > MIC	15 mg/kg every 6 hr SI (max: 1000 mg/dose)	1†	87% for institution 1
			4†	54% for institution 2
		25 mg/kg every 6 hr SI (max: 1000 mg/dose)	1†	94% for institution 1
			4†	57% for institution 2
Courter46 (N = 5000†; 2 yr, 12 yr; GP)	≥40% fT > MIC	15 mg/kg every 6 hr SI, EI3	2	45% SI, 95% EI
		25 mg/kg every 6 hr SI, EI3		55% SI, 99% EI
Doripenem
Matsuo83 (N = 5000†, 99§; 2 yr, 12 yr; GP)	≥40% fT > MIC	20 mg/kg every 8 hr SI, EI3 (max: 1000 mg/dose)	4	3% SI, 99% EI
		40 mg/kg every 8 hr SI, EI3 (max: 1000 mg/dose)	8	3% SI, 99% EI
Vaccaro84 (N = 20; 6–17 yr; CF)	fT > MIC	30 mg/kg EI4 (single dose) (max: 1000 mg/dose)	4	63% (Well tolerated by all subjects)
Zobell85 (N = 3; CF)	N/A	25–30 mg/kg every 8 hr EI4 (max: 1600 mg/dose)	N/A	(Improved pulmonary function tests. Mild ADR)

ADR, adverse drug reaction; CF, cystic fibrosis; CI, continuous infusion; CrCL, creatinine clearance; CSF, cerebrospinal fluid; EI, extended infusion; EI₃, extended infusion over 3 hrs; EI₄, extended infusion over 4 hrs; FEV1, forced expiratory volume in 1 second; FN, febrile neutropenia; GA, gestational age; GP, general pediatrics; HSCT; hematopoietic stem cell transplant; ICU, intensive care unit; LD, loading dose; LOS, length of stay; max, maximum; MIC, minimum inhibitory concentration; N/A, not applicable/available; NICU, neonatal ICU; PK-PD, pharmacokineticpharmacodynamic; PsA, Pseudomonas aeruginosa; PTA, probability of target attainment; SI, standard infusion

* Bolded regimens are the preferred regimens per study authors.

^† Median MIC reported.

^‡ Number of simulations.

§ Number of unique patients who contributed serum concentrations.

¶ Total number of patients who were retrospectively included in the study if they received standard or prolonged infusion piperacillin-tazobactam, cefepime, or meropenem for at least 72 hours.

# Clinical outcomes for individual antibiotic agents (piperacillin-tazobactam, cefepime, or meropenem) were not reported.

** units for creatinine clearance are mL/min/1.73 m²

Studies conducted by Padari et al⁷⁵ and Shabaan et al⁷⁶ have offered different levels of support for EI in neonates. Padari et al⁷⁵ compared steady-state PK for very low birth weight neonates treated with meropenem 20 mg/kg every 12 hours (SI, EI). As expected, a higher peak concentration was achieved with SI (89 mg/L) than EI (59 mg/L), while minimum serum drug concentrations consistently remained above the MIC of 2 mg/L in 8 of 10 patients. The authors concluded that benefits of EIs are less evident in very low birth weight neonates with prolonged meropenem half-lives. More favorable results were observed in a prospective trial by Shabaan et al⁷⁶ in which 102 neonates with Gram-negative, late-onset sepsis were randomly assigned to receive meropenem SI or EI at 20 mg/kg every 8 hours or 40 mg/kg every 8 hours for meningitis or pseudomonal infections. Neonates receiving EI had significantly higher rates of clinical improvement and microbiologic cure at 7 days, shorter duration of respiratory support, lower rates of mortality (31% SI; 14% EI; p = 0.03), and lower incidence of acute kidney injury (24% SI; 6% EI; p = 0.02) than those receiving SI. Application of this study is limited by lack of reported organism MICs.

In bacterial meningitis, data from 154 children was used to create a pediatric PK model including a cerebro-spinal fluid (CSF) compartment.⁷⁷ EIs did not significantly improve probability of attaining ≥50% fT > MIC(_CSF), but escalating the meropenem dose from 20 to 40 mg/kg every 8 hours notably increased PTA for penicillin-resistant Streptococcus pneumoniae and P aeruginosa. Lower peak concentrations and prolonged meropenem half-life in the CSF vs plasma may influence this finding. Clinical cure was achieved in 36 patients and was associated with ≥75% fT > MIC(_CSF).⁷⁷ In addition, a report has been published detailing an adolescent with intrathecal pump-related Achromobacter xylosoxidans meningitis who had treatment success with meropenem 40 mg/kg every 8 hours EI after SI failed without adverse events.⁷⁸ Although further study is needed, meropenem EI may be reasonable to consider in patients who do not improve with SI after maximizing the dose.

In the CF population, Pettit et al⁷⁹ developed a population PK model based on 30 children ages 6 to 17 years who received meropenem EI for pulmonary exacerbations. PTA to reach ≥40% fT > MIC for MICs of 1, 2, and 4 mg/L improved from <90% with SI to >99% for all 3 MICs with EI. Over half of patients completed their course with meropenem EI, including 5 who were discharged home on outpatient parenteral antibiotic therapy (OPAT) to complete treatment. Of the 12 patients who resumed SI, 9 had a regimen change to facilitate OPAT.⁷⁹ Complementing the previous PK study, Kuti and colleagues⁸⁰ investigated the association between achieving PK-PD parameters and positive clinical response. By analyzing 15 pediatric CF patients, the authors determined that achievement of fT > MIC was strongly associated with improvement in FEV1 (R² = 0.8, p = 0.0004). Furthermore, ≥65% fT > MIC was identified as a PD threshold for predicting improvement, with median FEV1 increase of 29% when this threshold was achieved vs 8% otherwise (p = 0.001).⁸⁰ The utility of meropenem CI in CF has been described in a case report detailing an adolescent with multi-drug resistant Inquilinus limosus (meropenem MIC = 4 mg/L). Meropenem dosing was converted to 3000 mg/day CI to achieve drug concentrations of 4 to 6 times the MIC after treatment failure with 500 mg every 8 hours SI (51 mg/kg/day).⁸¹ The dosing was later escalated to 6000 mg/day CI for an increased MIC of 8 mg/L. Pulmonary function tests improved with both CI regimens.⁸¹ Collectively, these data suggest meropenem EI may be preferred in hospitalized children with CF to achieve clinical response; however, EI during OPAT may be more difficult to integrate into a patient's home schedule.

Pediatric HSCT recipients treated with meropenem CI were studied to create a Monte Carlo simulation assessing 5 dosing regimens across 4 stratifications of renal function for severe Gram-negative infections.⁸² Doses of 45 to 60 mg/kg/day were needed to achieve 100% fT > 4× MIC in patients with augmented renal function.⁸² The aforementioned study by Zembles et al66 is also relevant to mention here, given their findings of decreased 30-day readmissions in HSCT patients receiving EI cefepime, piperacillin-tazobactam, or meropenem.

Imipenem/cilastatin. A Monte Carlo simulation of imipenem/cilastatin in children revealed that EI was necessary to achieve ≥90% PTA using a conventional PK-PD target at the CLSI P aeruginosa MIC breakpoint of 2 mg/L (Table 3).^46,67 This model also indicated that a MIC of ≤0.06 would be required for the standard pediatric dosing regimen of 60 mg/kg/day divided every 6 hours SI to achieve 90% PTA.⁴⁶ Neither microbiologic cure nor clinical outcomes for pediatric patients receiving EI imipenem/cilastatin have been described.

Doripenem. Doripenem EIs and CIs have been associated with improved clinical outcomes in critically ill adults, while pediatric data are limited to PK-PD modeling and small case series.^83–85 A Monte Carlo simulation developed by using serum concentrations from infants and young children demonstrated near 100% PTA when the doripenem dose was increased from 20 mg/kg to 40 mg/kg and administered as an EI.⁸³ In CF, a single dose of doripenem 30 mg/kg EI was administered to patients ages 6 to 17 years, with plasma concentrations resulting in AUC values consistent with those previously described in adult patients and 63% fT > MIC of 4 mg/L.⁸⁴ Doripenem was also associated with improvement in pulmonary function tests in 3 patients with CF treated with 75 to 90 mg/kg/day divided every 8 hours EI with minimal adverse effects.⁸⁵ Lastly, doripenem use has been reported in a critically ill child undergoing CRRT with suspected respiratory infection. Doripenem 60 mg/kg CI was predicted to achieve 100% fT > MIC for MIC ≤4 mg/L with CRRT clearance much higher than previous pediatric estimates (10.5 vs 4.4–4.8 mL/min/kg).⁸⁶

Monobactams

Aztreonam. Aztreonam EIs and CIs have been used successfully in small case series.⁸⁷ A 3-month-old infant with CF was treated with aztreonam 200 mg/kg/day CI plus tobramycin for 14 days, resulting in successful P aeruginosa eradication (MIC 4 mg/L) from respiratory cultures.⁸⁸ In a second case, a tracheostomy-dependent adolescent male was successfully treated for pneumonia involving 2 isolates of P aeruginosa (MIC 6 mg/L and 4 mg/L) with aztreonam 2 g every 6 hours EI in combination with polymyxin B. Aztreonam achieved 100% PTA in the serum and in the epithelial fluid (based on published lung concentrations relative to serum of 36%–80%) using defined goal of ≥40% fT > 4–6× MIC.⁸⁹ Lastly, an immunocompromised male was successfully treated for bilateral pneumonia and cutaneous ulcers growing multi-drug resistant P aeruginosa. Combination antibiotic therapy included aztreonam 6 to 8 g/day CI over 8 months and was well tolerated.⁹⁰ While data are limited, aztreonam EI is an alternative dosing strategy that may be used to treat multi-drug resistant Gram-negative infections.

Implementation Considerations

Drug stability is an important consideration when implementing EIs and CIs. Drug stability is dependent upon concentration, diluent, storage temperature, and infusion delivery device; while these factors apply to all IV preparations, temperature is one factor that requires careful assessment when prolonging infusions. Temperature of antibiotics delivered via CI may increase significantly if the OPAT portable infusion pump is worn under clothing and exposed to body temperature (37 °C) rather than room temperature (20 °C). Prescott et al⁹¹ have previously published a review of antipseudomonal beta-lactam stability. Authors conclude that piperacillintazobactam and aztreonam have reliable stability at body temperature, while ceftazidime and meropenem must be stored in cold environments during CI. Alternatively, the bags may be exchanged every 8 to 12 hours to prevent drug degradation, or in the case of ceftazidime, accumulation of the toxic metabolite pyridine.^91–93

Availability of uninterrupted IV access during prolonged infusions is another important consideration. While pediatric studies generally report low rates of resuming SI after EI, one of the most common reasons for discontinuing EI is inadequate IV access.^46,65 Patients may also need concomitant medications that are incompatible for coinfusion. On the other hand, SI nursing/caregiver burden and bacteremia risk associated with frequent IV catheter access may make use of EI and CI regimens attractive alternatives. For the home environment, caregiver preference regarding the practicality of each regimen should be considered to optimize OPAT adherence.⁹⁴

Risk of adverse events is another important consideration. Extended infusions and CIs are generally well tolerated, but infiltration of IV catheter has been reported in children.^{62,70,78,85,90,95} While beta-lactam TDM is beyond the scope of this review, PK-PD monitoring of peak, trough, or C_ss may guide patient-specific optimization of this class of drugs to achieve clinical efficacy and avoid toxicity.^22,24,96,97 An increasing body of literature demonstrates that beta-lactam TDM may have a role in clinical practice and is routinely recommended in critically ill adult patients by international groups.^22,24

Conclusion

PK-PD research has demonstrated that maintaining serum drug concentrations above the organism's MIC is essential to achieve bactericidal activity with beta-lactam antibiotics. Monte Carlo simulations have raised concerns that beta-lactam SI dosing fails to consistently achieve these targets in children, leaving patients vulnerable to poor response. Current literature suggests implementing beta-lactam EIs and CIs for pediatric patients is feasible and is associated with reliable PKPD target attainment along with similar or improved microbiologic and clinical responses. These regimens may be of particular benefit in several medically complex patient populations (CF, febrile neutropenia, critical care) and for infections with MICs at or near the clinical breakpoint. In children, the data for augmented infusion regimens is most extensive for piperacillin-tazobactam, meropenem, and ceftazidime (particularly in patients with CF). Evaluation of local MIC distributions, patient populations served, and nursing/caretaker preferences may guide which beta-lactam(s) to administer as EI or CI and whether the strategy should be standard for all children or those meeting specific criteria. Continuous infusion may also simplify OPAT regimens by decreasing the frequency of accessing the central line; however, consideration of drug stability and IV access is prudent.

ABBREVIATIONS

AUC

area under the curve

CF

cystic fibrosis

CI

continuous infusion

CLSI

Clinical and Laboratory Standards Institute

CRRT

continuous renal replacement therapy

CSF

cerebrospinal fluid

C_ss

steady-state concentration

ECMO

extracorporeal membrane oxygenation

EI

extended infusion

FEV1

forced expiratory volume in 1 second

FN

febrile neutropenia

fT > MIC

function of time free drug concentration remains above minimum inhibitory concentration

GA

gestational age

HSCT

hematopoietic stem cell transplant

ICU

intensive care unit

IV

intravenous

MIC

minimum inhibitory concentration

OPAT

outpatient parenteral antibiotic therapy

PD

pharmacodynamic

PK

pharmacokinetic

PTA

probability of target attainment

SI

standard infusion

TDM

therapeutic drug monitoring