Acute rheumatic fever (ARF), an autoimmune reaction to Group A Streptococcus (Streptococcus pyogenes; Strep A) infection, can cause rheumatic heart disease (RHD). New formulations of long-acting penicillins are being developed for secondary prophylaxis of ARF and RHD.

To evaluate the penicillin G concentrations required to suppress growth of Strep A.

Broth microdilution MIC and MBC for Strep A strains M75⁶¹¹⁰²⁴, M1T1⁵⁴⁴⁸ and M18^MGAS8232 were determined. All strains were studied in a hollow fibre model (initial inoculum 4 log₁₀ cfu/mL). Constant penicillin G concentrations of 0.008, 0.016 and 0.05 mg/L were examined against all strains, plus 0.012 mg/L against M18^MGAS8232. Viable counts were determined over 144 h. Subsequently, all penicillin G-treated cartridges were emptied, reinoculated with 5 log₁₀ cfu/mL and counts determined over a further 144 h. Mathematical modelling was performed.

MIC and MBC were 0.008 mg/L for all strains; small subpopulations of M75⁶¹¹⁰²⁴ and M1T1⁵⁴⁴⁸, but not M18^MGAS8232, grew at 1× MIC. Following the first inoculation, 0.008 mg/L achieved limited killing and/or stasis against M75⁶¹¹⁰²⁴ and M1T1⁵⁴⁴⁸, with subsequent growth to ∼6 log₁₀ cfu/mL. Following both inocula, concentrations ≥0.016 mg/L suppressed M75⁶¹¹⁰²⁴ and M1T1⁵⁴⁴⁸ to <1 log₁₀ cfu/mL from 6 h onwards with eradication. Concentrations ≥0.008 mg/L suppressed M18^MGAS8232 to <1 log₁₀ cfu/mL from 24 h onwards with eradication after both inoculations. Mathematical modelling well described all strains using a single set of parameter estimates, except for different maximum bacterial concentrations and proportions of bacteria growing at 1× MIC.

In the absence of validated animal and human challenge models, the study provides guidance on penicillin G target concentrations for development of new penicillin formulations.