Establishing the lowest penicillin concentration to prevent pharyngitis due to Streptococcus pyogenes using a human challenge model (CHIPS): a randomised, double-blind, placebo-controlled trial

Background: The in-vivo plasma concentration of penicillin needed to prevent Streptococcus pyogenes pharyngitis, recurrent acute rheumatic fever, and progressive rheumatic heart disease is not known. We used a human challenge model to assess the minimum penicillin concentration required to prevent streptococcal pharyngitis. Methods: In CHIPS, a randomised, double-blind, placebo-controlled, human challenge trial, healthy adult volunteers were randomly assigned by a computer-generated random sequence to target steady-state penicillin plasma concentrations (placebo, 3, 6, 9, 12, or 20 ng/mL). The study was a single-centre trial held in Perth, WA, Australia. Participants had to be healthy adults, aged 18–40 years, at low risk of complicated S pyogenes disease, and without high type-specific IgG antibodies against the emm75 S pyogenes challenge strain. Participants and staff involved in clinical care remained masked to treatment allocation for the duration of the study. Individualised 5-day continuous intravenous infusions of penicillin were commenced 12 h before direct pharyngeal application of the emm75 challenge strain. The primary endpoint was clinical pharyngitis. This trial is registered on the Australian New Zealand Clinical Trials Registry, ACTRN12621000751875, and is completed. Findings: Between Aug 23, 2022, and July 31, 2023, 60 participants were randomly assigned (35 [58%] were female and 25 [42%] were male), with 57 included in the analysis. The clinical pharyngitis endpoint was met in eight (57%) of 14 in the placebo group, four (44%) of nine in the 3 ng/mL target steady-state penicillin plasma concentration group, four (44%) of nine in the 6 ng/mL group, none of eight in the 9 ng/mL group, none of eight in the 12 ng/mL group, and none of nine in the 20 ng/mL group. No severe or serious adverse events occurred. Using Bayesian concentration–response modelling, the minimum steady-state plasma concentration of penicillin for which 90% of participants would avoid clinical pharyngitis was 8·1 ng/mL (95% credible interval 6·1–10·9). Interpretation: When steady-state penicillin concentrations are greater than 9 ng/mL, few people will develop experimental emm75 S pyogenes pharyngitis. These data will inform efforts to improve long-acting penicillin preparations and dosage regimens to prevent recurrent rheumatic fever and rheumatic heart disease.