Plasma nirmatrelvir and molnupiravir concentrations required for inhibition of SARS CoV-2 replication differ between rhesus macaques and humans

During the early COVID-19 pandemic, non-human primate (NHP) models, particularly rhesus macaques (RM), became vital for preclinical antiviral testing. However, their utility for predicting optimal human dosing remains uncertain. We applied previously validated mathematical models—originally fit to human clinical trial viral load data—to RM infected with SARS-CoV-2 and treated with nirmatrelvir/ritonavir, molnupiravir, or both. Our analysis revealed that both drugs exhibit greater antiviral potency in lungs than in nasal passages. Nirmatrelvir showed lower efficacy in RM (30% nasal, 46% lung) than in humans (82%), largely due to its shorter plasma half-life in macaques. In contrast, molnupiravir efficacy in RM (95% nasal, 99% lung) closely matched human efficacy against Omicron variants. Critically, achieving 50% viral inhibition in humans requires 10-fold higher nirmatrelvir plasma concentrations than in RM, while only 1/20th the molnupiravir concentration is needed in humans. These discrepancies stem from drug-specific differences in pharmacokinetics, pharmacodynamics, and tissue-specific viral dynamics between species. Consequently, while NHP models offer valuable insights into drug mechanisms and relative efficacy, they are insufficient for precise human dose optimization. Human Phase 1 and 2 clinical trial data remain superior for this purpose, as they directly reflect human physiology and drug response. Our findings caution against overreliance on NHP data for dosing predictions and underscore the need for human-centric modeling when translating preclinical results to clinical practice.