The dynamics of Plasmodium falciparum during the expansion phase of the asexual stage of infection

Abstract Background The asexual blood-stage expansion of Plasmodium falciparum is responsible for the onset of clinical malaria and is characterised by rapid parasite replication within the human host. Although within-host models have been widely used to study parasite dynamics, many fail to reproduce key features of observed parasite count trajectories, including the pronounced day-to-day fluctuations associated with parasite sequestration and synchrony, as well as the plateauing and decline in parasite densities observed during the expansion phase. Methods Daily parasite count data from 294 neurosyphilis patients infected with P. falciparum in historical malariotherapy studies were analysed together with 12 parasite trajectories from a contemporary controlled human malaria infection (CHMI) study measured by quantitative PCR (qPCR). An established age-structured model of parasite dynamics was extended to incorporate growth-limiting mechanisms acting on both merozoites and asexual blood-stage parasites. Models were fitted using a Bayesian framework with Markov chain Monte Carlo methods. Individual-level model comparison was performed using leave-one-out cross-validation, followed by population-level inference using a hierarchical model with strain-specific parameter distributions for three dominant parasite strains. Results Parasite count trajectories exhibited three characteristic patterns: sustained exponential growth, growth followed by a plateau, and growth followed by a decline. The extended model was favoured for 52% of patients, particularly those exhibiting non-exponential dynamics, and accurately reproduced both the timing and magnitude of growth attenuation. Population-level estimates indicated that growth-limiting effects typically became active 4–6 days after first microscopic detection and progressively reduced effective parasite multiplication. Initial parasite densities and age distributions were broadly similar across strains, although moderate differences in parasite multiplication rates and growth-limiting parameters were observed. For the qPCR data, both models reproduced the observed parasite trajectories and showed similar predictive performance. Conclusions Incorporating growth-limiting mechanisms into an age-structured model improves the ability to reproduce the diversity of within-host parasite dynamics observed during the expansion phase of P. falciparum infection. These findings highlight the importance of delayed and cumulative regulatory processes, potentially arising from host responses or density-dependent effects, in shaping parasite growth in vivo. The proposed framework provides a basis for future studies linking within-host parasite dynamics, immune responses, and therapeutic interventions.