Clinical considerations for oral beta-lactams as step-down therapy for Enterobacteriaceae bloodstream infections

KEYWORDS:

• Oral beta-lactam
• Enterobacteriaceae
• Gram-negative
• bacteremia
• bloodstream infections
• pharmacokinetics
• pharmacodynamics

1. Introduction

Oral antibiotic therapy for Enterobacteriaceae bloodstream infections (E-BSI) is becoming an increasingly appealing treatment strategy to avoid placement of indwelling catheter devices for antibiotic administration and their associated complications [1]. Historically, oral antibiotics such as fluoroquinolones and trimethoprim-sulfamethoxazole have been utilized for E-BSI based on supporting clinical and pharmacokinetic (PK)/pharmacodynamic (PD) data [1–4]. However, these antibiotics may not be optimal in specific clinical scenarios due to potential adverse effects associated with their use, including an increased risk for Clostridioides difficile-associated diarrhea with fluoroquinolones [1,5–7]. Due to the potential limitations of fluoroquinolones and trimethoprim-sulfamethoxazole, interest has spurred in the role of oral beta-lactams (OBL) as step-down therapy for E-BSI. However, there are important factors for clinicians to consider regarding their use for this indication.

2. Clinical data for OBL in Enterobacteriaceae bloodstream infections

Limited retrospective data exist evaluating the effectiveness of OBL as step-down therapy in E-BSI. These studies differ in the oral antibiotics evaluated, patient populations studied and defined outcomes (Table 1) [2–4]. As shown, there is considerable variability in the dosing strategies and specific OBL utilized in each of these studies. The majority of patients received amoxicillin-clavulanate and had urinary sources of infection with Escherichia coli being isolated most frequently.

Table 1. Summary of studies evaluating oral antibiotic regimens for Enterobacteriaceae bloodstream infections.

Study	Study design and population	Primary outcome	PO antibiotic studied	Most commonly reported dosing strategy of PO BL	Duration of therapy	Results
Kutob et al., 2016 [2]	Single-center, retrospective cohort study (n = 362)	Treatment failure rates of PO antibiotics for definitive therapy of GN BSI (all-cause mortality or recurrent infection within 90 days of initial episode)	High F (≥95%: levofloxacin [n = 106]) Moderate F (75–94%: ciprofloxacin [n = 151], TMP/SMX [n = 28]) Low F (penicillins [n = 42], cephalosporins [35])	A/C: 875/125 q12h (n = 21), 500/125 mg q8h (n = 9) Amoxicillin: 500 mg q8h (n = 10) Cephalexin: 500 mg q6h (n = 9) Cefuroxime: 500 mg q12h (n = 8) Cefdinir: 300 mg q12h (n = 7) Cefaclor: 250 mg q8h (n = 3) Cefpodoxime: 200 mg q12h (n = 1)	Mean 4.7 days of IV therapy followed by 9.1 days of PO therapy	Treatment failure risk greater with low F compared to high F: High F (2%); Moderate F (12%); Low F (14%); P = 0.02
Rieger et al., 2017 [4]	Single-center, retrospective cohort study (n = 241)	Treatment failure (escalation from PO to IV, antibiotic change, or 30-day readmission) of IV only versus partial or complete treatment with PO antibiotics for EB BSI from UTI source	Ciprofloxacin (63.5%) BL (19%) TMP/SMX (9.1%)	Not reported	Median IV therapy was 4 days (IQR 2–5) prior to PO switch	No statistically significant difference in treatment failure between IV only (3.8% [95% CI: 1.0–9.4%]) compared to IV/PO (8.2% [95% CI: 4.1 to 14.1%]; P = 0.19)
Mercuro et al., 2018 [3]	Single-center, retrospective cohort study (n = 224)	Compare efficacy (within 30 days after PO antibiotic completion) and safety between PO BL and FQ step-down therapy for EB BSI	BL (n = 84): A/C (n = 44), amoxicillin (n = 12), cephalexin (n = 19), cefdinir (n = 9); FQ (n = 140): ciprofloxacin (n = 84), levofloxacin (n = 55)	A/C: 875/125 mg q12h (n = 31), 500/125 mg q12h (n = 9), 500/125 mg q8h (n = 3), 500/125 mg q24h (n = 1); amoxicillin: 1000 mg q8h (n = 4), 1000 mg q12h (n = 2), 500 mg q8h (n = 3) 500 mg q12h (n = 3); cephalexin: 500 mg q6h (n = 11), 500 mg q8h (n = 6), 500 mg q12h (n = 2); cefdinir: NR	Median IV therapy was 3 days (IQR 3–4) Median total duration of therapy: 14 days (IQR 11–15)	PO BL was non-inferior to PO FQ (86.9% vs. 87.1%; mean difference 0.2%, 97.5% CI −10.3 to 10.7). Microbiological success (94.0% vs. 97.9%; P > 0.05) and 30-day re-admission (14.3% vs. 14.3%; P > 0.05) were similar

Abbreviations: A/C: amoxicillin-clavulanate; BL: beta-lactams; BSI: bloodstream infections; EB: Enterobacteriaceae; F: bioavailability; FQ: fluoroquinolone; GN: Gram-negative; IV: intravenous; NR: not reported; PO: oral; qxh: every x hours; TMP/SMX: trimethoprim-sulfamethoxazole; UTI: urinary tract infections.

Kutob and colleagues observed that patients receiving low bioavailability agents (any OBL) were 7.67 times more likely to experience treatment failure compared to patients receiving a high bioavailability agent (levofloxacin) [2]. This study included patients receiving seven different OBL with various dosing regimens and categorized all beta-lactams as low bioavailability agents despite significant variability in bioavailabilities between the agents used (Table 2). Additional factors including protein binding and volume of distribution were not considered in the categorization of the OBL that may considerably impact the likelihood of PD target attainment between different agents. Lastly, although the majority of isolates from this study were E. coli, patients infected with other bacteria including Pseudomonas aeruginosa and chromosomally mediated Amp-C producing Enterobacteriaceae were included, which may have also resulted in potential selection bias for antibiotic selection and influenced the observed effectiveness of OBL.

Table 2. Pharmacokinetic parameters of commonly used oral penicillins and cephalosporins [8,9].

Antibiotic Class	Antibiotic	Bioavailability (%)	Protein binding (%)	Food Effect on Absorption	Peak Serum Concentration (mg/L)	Half-life (hours)
Oral Penicillins	Amoxicillin	74–92	20	Not Significant	3.5–5.0	1.0
	Amoxicillin- clavulanate	60	20	Not Significant	3.7–4.8	1.0
Oral Cephalosporins	Cephalexin	90–100	6–15	Not Significant	15–18	0.6–1.3
	Cefaclor	52–95	25	Not Significant	13–15	0.6–1.0
	Cefprozil	71–95	36	Not Significant	8.2–10.4	0.9–1.4
	Cefuroxime	30–52	33–50	Increase	7.0	1.0–2.0
	Cefdinir	21–25	60–70	Not Significant	1.6–2.3	1.7
	Cefpodoxime	29–53	22–33	Increase	3.9–4.5	2.2–2.8

Mercuro and colleagues found OBL to be non-inferior to fluoroquinolones with regards to clinical success when used as step-down therapy in E-BSI [3]. This study included patients who received amoxicillin-clavulanate, amoxicillin, or cephalexin with at least three different dosing regimens that were further stratified by creatinine clearance. Patients received a median of 3 days of intravenous antibiotics prior to oral transition. Evaluation of clinical success, defined as completion of therapy without new onset sepsis after oral step-down therapy, all-cause mortality at 30 days, or documented change in antibiotic therapy due to an adverse event or clinical worsening, is limited by the single-center, retrospective design and potential loss of follow-up or incomplete data, especially considering that more than half of the patients were transitioned to an oral agent on the day of discharge. However, patients were more likely to complete their antibiotic course without an adverse effect with an OBL compared to a fluoroquinolone (91.7% vs. 82.1%; P = 0.049).

Lastly, Reiger and colleagues found no statistically significant difference in treatment failure between patients who received intravenous-only compared to intravenous followed by oral antibiotics for E-BSI, specifically from a urinary source [4]. However, there were a limited number of patients who received OBL, and dosing regimens were not provided, making interpretation of these results for specific OBL challenging. Based on the available clinical data with conflicting results and variations in OBL selection and dosing regimens, drawing inferences on the effectiveness of OBL as a class is challenging.

3. Pharmacokinetic and pharmacodynamic considerations

OBL consist of both penicillins and cephalosporins which exhibit notable pharmacokinetic differences that distinguish specific agents from one another (Table 2) [1–4,8,9]. Despite often being labeled as low bioavailability agents, the bioavailabilities of specific OBL vary considerably. For example, the bioavailability of cephalexin actually exceeds that of ciprofloxacin (~70%), which is often labeled as having high or moderate bioavailability and is much higher than that of amoxicillin-clavulanate or cefdinir [8–10].

While bioavailability is an essential PK parameter as it relates to serum and infection site drug concentrations, serum protein binding and volume of distribution may be of equal importance to determine if an antibiotic is able to meet or exceed PD targets [11]. Beta-lactam antibiotics exhibit time-dependent pharmacodynamics, yet the recommended time that free drug concentrations remain above the minimum inhibitory concentration (fT>MIC) vary between beta-lactam classes. It is generally recommended to maintain fT>MIC for 50% and 60–70% of the dosing interval for penicillins and cephalosporins, respectively [12]. Table 3 displays the estimated percentage of fT>MIC for various OBL at specific MIC values as well as the maximum MIC value that would allow for PD target attainment. However, determination of MIC values for OBL is often not routinely available, which poses an additional challenge.

Table 3. Percentage of free time above the minimum inhibitory concentration for various oral beta-lactam antibiotics.

Antibiotic	Dose (mg)/dosing interval (h)	%fT>MIC							Maximum MIC allowing for target attainment	Highest frequency wild-type MIC for E. coli*
		16 mg/L	8 mg/L	4 mg/L	2 mg/L	1 mg/L	0.5 mg/L	0.25 mg/L
Amoxicillin	500/8	–	13.0	23.0	33.0	43.0	53.0	63.0	0.5 mg/L	4 mg/L
Amoxicillin	1000/8	–	23.0	33.0	43.0	53.0	63.0	73.0	1 mg/L	4 mg/L
Amoxicillin-clavulanate	875/12	–	11.0	17.6	24.3	31.0	37.6	44.3	–	4 mg/L
Amoxicillin-clavulanate	875/8	–	16.4	26.4	36.4	46.4	56.4	66.4	0.5 mg/L	4 mg/L
Cephalexin	500/6	3.30	22.7	42.1	61.5	80.9	100	100	2 mg/L	4 mg/L
Cephalexin	1000/6	22.7	42.1	61.5	80.9	100	100	100	4 mg/L	4 mg/L
Cefaclor	500/6	–	11.0	23.5	36.0	48.5	61.0	73.5	0.5 mg/L	1 mg/L
Cefprozil	500/12	–	12.6	20.1	27.5	35.0	42.5	50.0	–	N/A
Cefuroxime	500/12	–	–	0.43	10.2	20.5	29.7	39.5	–	4 mg/L
Cefdinir	300/12	–	–	–	–	7.41	12.4	17.3	–	N/A
Cefpodoxime	400/12	–	–	–	22.8	39.7	56.6	73.5	0.25 mg/L	0.5 mg/L

Calculations are made according to a one-compartment first-order elimination model adapted from Turnidge JD [13] and are based on the mean bioavailability, mean protein binding percentage, and longest half-lives reported for each agent from Table 2. Calculated %fT>MIC is based on a 70-kg person and does not account for absorption time. Target attainment is based on a %fT>MIC of at least 50% and at least 60% for penicillins and cephalosporins, respectively. See Supplemental Material for calculations. *Based on the Antimicrobial Wild Type Distributions of Microorganisms from the European Committee on Antimicrobial Susceptibility Testing [16]. N/A; not available.

Abbreviations: %fT>MIC: percentage of time free drug concentrations remain above the minimum inhibitory concentration; MIC: minimum inhibitory concentration

4. Susceptibility testing, agent selection, and dosing considerations

Susceptibility testing of OBL is not routinely performed across many institutions. Instead, cefazolin susceptibility may be used as a surrogate to predict susceptibility for specific oral cephalosporins, but only for uncomplicated urinary tract infections (UTI) caused by E. coli, Klebsiella pneumoniae, and Proteus mirabilis [14]. In clinical practice, clinicians will often use cefazolin susceptibility to predict the susceptibility of oral cephalosporins for infectious indications other than UTI as well. Additionally, the results of ampicillin testing may be used to predict susceptibility for amoxicillin. Nevertheless, utilization of a surrogate to predict susceptibility still does not provide a tangible MIC to allow for assessment of PD target attainment.

In the absence of specific MIC values, the use of wild-type MIC distributions may be considered when selecting an OBL that is predicted to be susceptible based on surrogate susceptibility testing [15]. For example, wild-type MIC values occurring in the highest frequency for amoxicillin, amoxicillin-clavulanate, cephalexin, cefaclor, cefuroxime, and cefpodoxime for E. coli are 4 mg/L, 4 mg/L, 4 mg/L, 1 mg/L, 4 mg/L, and 0.5 mg/L, respectively [16]. Taking these values into consideration with the data presented in Table 3, cephalexin 1000 mg every 6 h appears to offer the highest likelihood of PD target attainment amongst OBL in a patient with normal renal function, however, amoxicillin, amoxicillin-clavulanate (when dosed every 8 h), cefaclor, and cefpodoxime may be suitable options at lower MICs, if known. Cefprozil, cefuroxime, and cefdinir appear to offer a low likelihood of PD target attainment. In patients with renal impairment requiring OBL dose adjustments, the overall exposure and influence on PD target attainment are unclear.

5. Duration of therapy for enterobacteriaceae bloodstream infections

Two final considerations are the overall length of therapy and the duration of intravenous antibiotics prior to oral step-down therapy for E-BSI. In observational studies, patients typically received between 3 and 5 days of intravenous therapy prior to transitioning to an OBL to complete a total antibiotic course of approximately 14 days [2–4]. A recent randomized controlled trial demonstrated non-inferiority of 7 days of antibiotics compared to 14 days for uncomplicated E-BSI [17]. This study included patients who received an OBL as step-down therapy, however, only a minority of patients (14.3% and 20.7% in the 7-day and 14-day arms, respectively) were transitioned to an OBL. Furthermore, specific outcomes for patients receiving an OBL for step-down therapy are not reported. While shorter antibiotic courses are advisable in many clinical scenarios for E-BSI, longer courses (>7 days) may still be warranted when transitioning to an OBL as step-down therapy.

6. Expert opinion

OBL are being increasingly considered as step-down therapy for E-BSI given their tolerability profile compared to fluoroquinolones and trimethoprim-sulfamethoxazole. However, several important concerns remain that preclude routine use.

Unlike intravenous beta-lactams, the dose of OBL required to attain specific PD targets is unclear, especially when MIC values are commonly unknown. This is further complicated when dose adjustments are needed for patients with renal impairment. Additionally, given the short half-life of many OBL, frequent dosing is often required which may impact patient adherence. Lastly, the heterogeneity of limited available studies evaluating OBL makes it difficult to interpret their clinical efficacy as step-down therapy in E-BSI.

As such, OBL should not be considered as preferred step-down therapy for E-BSI. However, these agents may be considered in specific clinical scenarios where the risks outweigh the benefits of oral fluoroquinolones, oral trimethoprim-sulfamethoxazole, or continuation of intravenous antibiotics. Based on available data and clinical expertise, OBL should only be considered in immunocompetent patients with a presumed low inoculum source of infection and functional gastrointestinal tract after completion of at least 3 days of intravenous antibiotics. If this strategy is employed, the specific agent and dose selected should be one expected to achieve pharmacodynamic target attainment. Based on wild-type MIC distributions, cephalexin 1000 mg every 6 h appears to offer the highest likelihood of PD target attainment amongst OBL, however, at lower MICs, amoxicillin, amoxicillin-clavulanate (dosed every 8 h), cefaclor, and cefpodoxime may be appropriate. Despite data supporting shorter antibiotic courses for E-BSI treatment, a total duration of 10 to 14 days may still be advisable when utilizing OBL step-down therapy. Additional clinical and PK/PD data are needed to further evaluate the role and optimal dosing strategies of OBL for E-BSI.

Declaration of interest

WD Kufel has served on an advisory board of Theratechnologies and has received a research grant from Melinta Therapeutics while JM Steele has served on an advisory board for Paratek Pharmaceuticals. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.

Reviewer disclosures

Peer reviewers on this manuscript have no relevant financial or other relationships to disclose.

Supplementary material

Supplemental data for this article can be accessed here.

Additional information

Funding

This manuscript was not funded.