Population pharmacokinetics of penicillin G: insights into increased clearance at low concentrations to guide development of improved long-acting formulations for syphilis and prevention of rheumatic fever

Abstract

Although benzylpenicillin (penicillin G) is listed by the World Health Organization as an Essential Medicine, dose optimization is a persistent challenge, especially for long-acting intramuscular formulations. Maintaining sustained antibiotic exposure at target concentrations is crucial for secondary chemoprophylaxis of rheumatic heart disease and treatment of syphilis. This study compared the pharmacokinetic profile of continuous low-dose benzylpenicillin infusions with a standard-dose bolus and evaluated which renal function marker (serum creatinine, cystatin C, or combined e-glomerular filtration rate [eGFR]) best predicted clearance. Healthy adult volunteers received a single 600 mg IV benzylpenicillin bolus followed by randomization to continuous infusions targeting steady-state concentrations of 3, 6, 9, 12, or 20 ng/mL. Plasma benzylpenicillin concentrations were measured by liquid chromatography-mass spectrometry. Population pharmacokinetic analysis was performed using NONMEM by incorporating both bolus and infusion data, and various GFR estimations were evaluated as covariates for clearance. Data from 72 participants were analyzed, including 504 bolus and 389 continuous infusion samples. A two-compartment model improved fit when the ratio of central volume of distribution between bolus and low-dose infusion was incorporated, and clearance differences at steady state plasma concentration of 3 ng/mL were accounted for. Of the GFR estimations, cystatin C-based eGFR significantly enhanced model fit compared with creatinine-based equations. Benzylpenicillin pharmacokinetics at very low concentrations demonstrated both a higher volume of distribution and increased clearance. Cystatin C-based eGFR may more accurately predict benzylpenicillin clearance, enabling precision dosing for long-acting preparations used for treatment of syphilis and prevention of rheumatic fever.

Keywords: benzathine penicillin G; benzylpenicillin; cystatin C; eGFR; penicillin G; pharmacokinetics; population PK modeling; renal clearance; rheumatic fever; rheumatic heart disease; syphilis.

Fig 1

Discrepancies between benzylpenicillin clearance (L/h) estimates from bolus dose (CL_FO) and continuous infusion (CL_inf), particularly at lower CL_FO. The normalized steady-state clearance (CL_inf/CL_FO) is calculated by dividing the continuous-infusion clearance by the bolus-based clearance. Points are colored by target steady-state concentration group (3, 6, 9, 12, and 20 ng/mL).

Fig 2

Relationship between infusion rate (RATE, µg/h) and benzylpenicillin clearance (L/h). (A) Steady-state clearance (CL_inf) vs. RATE. (B) Clearance estimated from a single intravenous bolus 600 mg dose (CL_FO) vs. RATE. (C) Normalized steady-state clearance (CL_inf/CL_FO) for rates below 1200 vs. RATE. (D) Normalized steady-state clearance (CL_inf/CL_FO) for rates below 100 m vs. RATE. Points are colored by target steady-state concentration group (3, 6, 9, 12, and 20 ng/mL). Data are from the CHIPS trial, a double-blind, placebo-controlled, randomized trial of benzylpenicillin for the prevention of experimental pharyngitis.

Fig 3

Diagnostic plots for the population pharmacokinetic model of bolus infusion. (A) Observed versus population-predicted benzylpenicillin concentrations, (B) Observed versus individual-predicted benzylpenicillin concentrations, (C) conditional weighted residuals versus time, and (D) conditional weighted residuals versus population-predicted concentrations. The dashed lines represent the lines of identity, whereas the solid lines indicate the lines of best fit.

Fig 4

Visual predictive checks for benzylpenicillin plasma concentrations following a 600 mg bolus infusion and subsequent low-dose continuous infusions. (A) displays data from bolus doses only. (B–F) include both bolus and low-dose infusion data and correspond to target steady-state concentrations of 3, 6, 9, 12, and 20  ng/mL, respectively. Open circles represent the measured plasma concentrations of benzylpenicillin. Solid lines denote the median values of the observed data, whereas dashed lines indicate the 5th and 95th percentiles of the observed data. Shaded areas illustrate the 95% confidence intervals for the simulated values generated by the pharmacokinetic model, with the upper and lower shaded areas corresponding to the 95th and 5th percentiles, respectively, and the middle shaded area representing the 50th percentile.

Fig 5

Diagnostic plots for the population pharmacokinetic model of low dose continuous infusion. (A) Observed versus population-predicted benzylpenicillin concentrations, (B) Observed versus individual-predicted benzylpenicillin concentrations, (C) Conditional weighted residuals versus time, and (D) Conditional weighted residuals versus population-predicted concentrations. The dashed lines represent the lines of identity, whereas the solid lines indicate the lines of best fit. The low-dose continuous infusion rate ranged from 46.9 to 1133.6  µg/h.