Optimal Interruption of P. vivax Malaria Transmission Using Mass Drug Administration

Abstract

Plasmodium vivax is the most geographically widespread malaria-causing parasite resulting in significant associated global morbidity and mortality. One of the factors driving this widespread phenomenon is the ability of the parasites to remain dormant in the liver. Known as ‘hypnozoites’, they reside in the liver following an initial exposure, before activating later to cause further infections, referred to as ‘relapses’. As around 79–96% of infections are attributed to relapses from activating hypnozoites, we expect it will be highly impactful to apply treatment to target the hypnozoite reservoir (i.e. the collection of dormant parasites) to eliminate P. vivax. Treatment with radical cure, for example tafenoquine or primaquine, to target the hypnozoite reservoir is a potential tool to control and/or eliminate P. vivax. We have developed a deterministic multiscale mathematical model as a system of integro-differential equations that captures the complex dynamics of P. vivax hypnozoites and the effect of hypnozoite relapse on disease transmission. Here, we use our multiscale model to study the anticipated effect of radical cure treatment administered via a mass drug administration (MDA) program. We implement multiple rounds of MDA with a fixed interval between rounds, starting from different steady-state disease prevalences. We then construct an optimisation model with three different objective functions motivated on a public health basis to obtain the optimal MDA interval. We also incorporate mosquito seasonality in our model to study its effect on the optimal treatment regime. We find that the effect of MDA interventions is temporary and depends on the pre-intervention disease prevalence (and choice of model parameters) as well as the number of MDA rounds under consideration. The optimal interval between MDA rounds also depends on the objective (combinations of expected intervention outcomes). We find radical cure alone may not be enough to lead to P. vivax elimination under our mathematical model (and choice of model parameters) since the prevalence of infection eventually returns to pre-MDA levels.

Mathematical models of Plasmodium vivax transmission: A scoping review

“…Out of the 48 studies identified that incorporated a P. vivax transmission model, 38 (79%) utilised a deterministic and differential equation (compartmental) framework [ 4 , 20 , 24 , 28 – 30 , 32 – 35 , 38 , 39 , 56 , 58 – 82 ] and nine (19%) used a stochastic and agent-based framework [ 21 , 54 , 55 , 83 – 88 ] ( Fig 2 ). Only one study (2%) used both deterministic and stochastic frameworks [ 20 ] to model P. vivax transmission.…”

“…Only a handful of models account for the variability in hypnozoite inoculation across mosquito bites (i.e. hypnozoite density model, boxes with a bright pink border in Fig 2 ) [ 4 , 32 , 34 , 35 ]. If the size of the hypnozoite reservoir is modelled explicitly, the number of compartments in the model increases substantially.…”

“…Depending on the setting the model is being developed for, these relapse patterns may also need to be taken into account. In most of the models, the tropical relapse pattern is achieved with a shorter latent period [ 4 , 20 , 24 , 28 – 30 , 35 , 38 , 39 , 56 , 58 – 62 , 64 – 66 , 68 – 72 , 74 – 76 , 78 – 82 ]. Only a few models have considered both temperate and tropical relapse patterns [ 32 – 34 ].…”

“…Generally, there are two approaches when incorporating superinfection: (i) using a corrected recovery rate that explicitly accounts for the history of past infections in the population and hypnozoite accrual dynamics [ 4 , 36 , 113 ] and (ii) coupling the prevalence of blood-stage infection (derived under a within-host model that accounts for superinfection) directly to the proportion of infected mosquitoes [ 36 ]. The within-host model of Mehra et al .…”

“…The impact of radical cure treatment on the infected population has also been studied in models [ 21 , 30 , 32 , 64 , 85 ] where only a handful of models incorporate G6PD testing [ 21 , 64 , 85 ]. Furthermore, 11 studies (23%) evaluated the effect of MDA on disease transmission, despite few national control programs considering MDA for P. vivax control [ 4 , 20 , 28 , 29 , 34 , 54 – 56 , 72 , 87 , 88 ] (boxes with a dark green border in Fig 2 ). Since MDA is recommended as an important tool to reduce asymptomatic P. falciparum infection, it is also likely to be of great importance for P. vivax elimination [ 20 , 146 , 147 ].…”