Non-Linear Absorption Pharmacokinetics of Amoxicillin: Consequences for Dosing Regimens and Clinical Breakpoints

Amoxicillin is an aminopenicillin that has been used to treat bacterial infections since the 1970s.

It is frequently combined with the b-lactamase inhibitor clavulanic acid to target b-lactamase producing strains. Either alone or in combination with clavulanic acid, amoxicillin was the most consumed antibacterial agent in primary care in two-thirds of the EU/EEA countries in 2012.

In an era of increasing antimicrobial resistance and with few new drugs making it to the market, antibiotic use must be optimized in order to improve clinical outcomes of infections. These clinical outcomes are dependent on the relationship between MIC, efficacy and exposure. For b-lactams, the pharmacokinetic/pharmacodynamic (PK/PD) index that best correlates with efficacy is the duration that the unbound concentration exceeds the MIC, expressed as a percentage of the dosing interval (%fT>MIC).

The minimal PK/PD index value that ensures a high probability of successful treatment is the PD target. The PD target appears to be different for each b-lactam group (penicillins, cephalosporins and carbapenems). For penicillins, the PD target is >30%–50% fT>MIC, dependent on the microbial species and the choice of the antibacterial endpoint (i.e. the necessary %fT>MIC is larger for a 1 or 2 log bacterial kill than for a static effect). While a high %fT>MIC is related to increased bacterial efficacy, an inadequate %fT>MIC is associated with emergence of resistance and selection of resistant strains. To attain a specific PD target, the exposure of the microorganism to the antibacterial agent needs to be adequate. This exposure is dependent on the dose and PK properties of the drug.

Despite the abundant use of oral amoxicillin, the drug’s PK has been described in only a few studies. While small-scale PK studies have shown amoxicillin to have a non-linear absorption profile, it remains unclear how such non-linear absorptionmight influence the exposure of the various dosing regimens. At present, standard dosing regimens of oral amoxicillin in adults and children ≥40 kg vary between 750 and 3000 mg/day, divided into two to four doses (e.g. 250, 500 or 1000 mg three times daily, 500 mg twice daily and 500 mg four times daily). For oral amoxicillin/clavulanic acid, standard dosing regimens are 500/125 mg three times daily or 875/125 mg twice or three times daily.

A population PK model can be used to estimate the exposure of various dosing regimens and variability of the antibiotic in the population. However, such a model is currently not available for oral amoxicillin in the literature. Monte Carlo simulations based on a population PK model can support recommendations for more-appropriate dosing regimens with a reduced likelihood of ineffectiveness and resistance (with too low doses) and adverse events (with too high doses). Information about the PD target, PK, exposure, variability and dosing regimens is needed to set clinical breakpoints. Clinical breakpoints are MICs that define microorganisms as susceptible, intermediate or resistant to specific antibiotics.4 The purposes of this study were to estimate the exposure of various oral amoxicillin dosing regimens and the variability in the population, to compare the probability of PD target attainment of these dosing regimens and to suggest which clinical breakpoints would be justified for oral dosing.We therefore developed a population PK model using NONMEM and performed Monte Carlo simulations.