Modelling the contribution of relapse infections to the epidemiology and control of Plasmodium vivax malaria

Plasmodium vivax is the most widely distributed species of malaria across the world, with almost 3 billion people living in at-risk countries, and an estimated 100 - 400 million clinical cases every year, mostly in Asia and South America. Despite the enormous burden to public health, research into P. vivax is neglected compared to P. falciparum (the parasite species responsible for the majority of cases of malaria in Africa). A key feature of P. vivax malaria not seen in all forms of malaria is the occurrence of relapse infections, where infected individuals become re-infected weeks to months after recovering from their initial infection. This is a phenomenon occasionally seen in British and European travellers returning from Asia and South America. Despite taking daily anti-malarial tablets during their trip, some travellers suffer relapse infections upon their return and may experience a potentially fatal episode of clinical malaria, often up to a year later. In contrast, relapse infections are a regular occurrence for people living in countries with P. vivax malaria. However in practice, in regions of the world where P. vivax is continuously transmitted it is difficult to distinguish these relapses from new infections arising from mosquito bites.

Relapse infections arise from P. vivax parasites that lie dormant in the liver and re-activate at a later date. The triggers for relapse infections are still poorly understood, with explanations ranging from fever due to other diseases, exposure to subsequent mosquito bites, and a biological clock for relapse infections. In this project, I aim to explore what causes infections to relapse by using simulation models to analyse data from approximately 4,000 people exposed to P. vivax infection in trials in Papua New Guinea, the Solomon Islands, Thailand and Brazil. With an understanding of how infections relapse, I will develop a simulation model of how P. vivax is transmitted between humans and mosquitoes, and use this to investigate how scaling up malaria control interventions can be used to control or eliminate P. vivax from endemic countries. P. vivax can be prevented by using insecticide treated nets to kill infected mosquitoes, or anti-malarial drugs to kill the parasites inside humans. Most anti-malarial drugs only target blood-stage infection and leave parasites in the liver untouched. A notable exception is primaquine which targets both parasites in the blood and dormant parasites in the liver.

A century of experience of malaria control has shown that drug treatment and vector control reduce cases of P. vivax malaria, however the impact of these interventions on population level transmission has been difficult to predict. I will use a simulation model of the transmission of P. vivax to compare how treatment programmes with drugs targeting different stages in the parasite's lifecycle can protect the infected individual, and the wider community by preventing onward transmission. Insights from these models will contribute to the evidence base required to choose optimal combinations of malaria control interventions as national malaria control programmes strive for increased P. vivax control and in some cases, elimination.

Following a bite from a P. vivax infectious mosquito, sporozoites inoculated into the skin migrate to the liver, where they develop to cause blood-stage infection, or transform into hypnozoites, a stage of the parasite that lies dormant in the liver. Hypnozoites re-activate weeks to months later and cause P. vivax relapses. Recent advances in genotyping markers for molecular monitoring have enabled investigation of the genetic diversity and multiplicity of P. vivax infections, allowing relapses to be distinguished from new infections. I will use data on genotyped infections from approximately 4,000 participants in longitudinal studies to estimate the contributions of new and relapse infections. These estimates will be combined with epidemiological data to test hypotheses for the determinants of hypnozoite activation, ranging from febrile illness, exposure to Anopheles proteins, to a seasonally-driven biological clock for activation.

I will construct statistical models of P. vivax relapses and fit them to data from treatment-reinfection studies to estimate key parameters describing relapses: (i) the proportion of infections due to relapses; (ii) the frequency of relapse infections; and (iii) the period during which a person is at risk of relapsing after primary infection. Models of relapse infections will be embedded within a transmission model of the P. vivax lifecycle. This will be fitted to data on age and exposure-specific patterns of parasite prevalence and clinical incidence using Bayesian Markov Chain Monte Carlo methods incorporating prior biological knowledge. Using this model, I will estimate the public health benefit of malaria control interventions. I will develop models for the effects of primaquine treatment (the only available drug that can clear hypnozoites), to compare different treatment strategies, from the perspective of an individual receiving treatment, and from a population-level perspective incorporating reductions in transmission.

The proposed research has the potential to benefit both academic and non-academic sectors. Below I outline the steps that will be taken to ensure that the research has the maximum impact in a number of distinct groups.

1. Academic research groups

The findings of this research are likely to be of interest to many academic research groups. In particular the validation and/or refutation of hypotheses for the biological determinants of relapse infections, a deeper understanding of the transmission dynamics of P. vivax, and estimates of the public health benefits of combinations of malaria control interventions.

2. Malaria control policy makers

The lack of capacity in the mathematical modelling of P. vivax malaria has repeatedly been highlighted as an obstacle to malaria control and elimination efforts, for example at the Advances in Plasmodium vivax Research conference, the modelling section of the Research Agenda for Malaria Eradication published recently in PLoS Medicine, and the Malaria Policy Advisory Committee's recommendations for the WHO's Elimination Scenario Planning Tool. A key objective of this project is to produce an easy to use statistically validated mathematical model of P. vivax transmission capable of providing quantitative estimates of the public health benefits of scaling up malaria control interventions.

3. Local malaria control programme managers

A key output of the work of the Malaria research group at Imperial College London was a software tool to allow local malaria control programme managers to estimate the effects of interventions such as insecticide treated nets and increased treatment of P. falciparum cases. This software was developed in collaboration with the World Health Organisation and has been successfully trialled with Malaria Control Programme officers in Senegal and The Gambia. In the long term, I would hope to take advantage of the existing technical capacity at Imperial College to deliver a similar resource for P. vivax control and elimination efforts.

4. Pharmaceutical companies

The use of mathematical models to estimate the effectiveness treating P. vivax malaria on both the individual and population-level, has the potential to benefit any parties involved in the manufacture and distribution of anti-malarial drugs, from pharmaceutical companies to non-profit foundations such as Medicines for Malaria Venture. The only licensed drug that can provide radical cure of P. vivax liver-stage infections is primaquine - a generic drug manufactured by several companies. A primaquine analogue, tafenoquine, is being developed by GlaxoSmithKline (GSK), and is currently undergoing Phase 3 trials. Tafenoquine could potentially become licensed for use in 2017. Past experience of collaborating with partners from industry puts me in a strong position to tailor output from research (academic papers and other scientific reports) so as to have maximum impact.


5. Members of the public in the UK

Travellers to malaria endemic countries from the UK usually receive sound medical advice on whether or not anti-malarial chemoprophylaxis is needed. However the differences in risk after exposure to Plasmodium vivax or Plasmodium falciparum are not usually perceived by members of the public. In particular, the fact that P. vivax can cause relapse infections months after returning home, even if anti-malarials were taken, is not appreciated by many travellers.

6. Individuals living in P. vivax endemic countries

The ultimate beneficiaries of this research should be individuals living in P. vivax endemic countries. Malaria control and elimination efforts require enormous levels of cooperation and collaboration from across a broad range of partners. Mathematical models have a small but vital role to play in these efforts, ensuring that health benefits to populations living in endemic countries are accrued over the fastest possible time scale.