Estimating Geographical Variation in the Risk of Zoonotic Plasmodium knowlesi Infection in Countries Eliminating Malaria

Abstract

Background: Infection by the simian malaria parasite, Plasmodium knowlesi, can lead to severe and fatal disease in humans, and is the most common cause of malaria in parts of Malaysia. Despite being a serious public health concern, the geographical distribution of P. knowlesi malaria risk is poorly understood because the parasite is often misidentified as one of the human malarias. Human cases have been confirmed in at least nine Southeast Asian countries, many of which are making progress towards eliminating the human malarias. Understanding the geographical distribution of P. knowlesi is important for identifying areas where malaria transmission will continue after the human malarias have been eliminated.

Methodology/principal findings: A total of 439 records of P. knowlesi infections in humans, macaque reservoir and vector species were collated. To predict spatial variation in disease risk, a model was fitted using records from countries where the infection data coverage is high. Predictions were then made throughout Southeast Asia, including regions where infection data are sparse. The resulting map predicts areas of high risk for P. knowlesi infection in a number of countries that are forecast to be malaria-free by 2025 (Malaysia, Cambodia, Thailand and Vietnam) as well as countries projected to be eliminating malaria (Myanmar, Laos, Indonesia and the Philippines).

Conclusions/significance: We have produced the first map of P. knowlesi malaria risk, at a fine-scale resolution, to identify priority areas for surveillance based on regions with sparse data and high estimated risk. Our map provides an initial evidence base to better understand the spatial distribution of this disease and its potential wider contribution to malaria incidence. Considering malaria elimination goals, areas for prioritised surveillance are identified.

Fig 1. Schematic overview of the methods.

Blue boxes describe input data, green boxes denote analyses, and yellow boxes represent final outputs. MBS = Malaysia, Brunei and Singapore.

Fig 2. Occurrence data used for model fitting and evaluation.

A. Location of presence and absence points/polygons outside Malaysia, Brunei and Singapore used for model evaluation. B. Location of presence and absence points/polygons as well as background points from Malaysia, Brunei and Singapore used for model fitting.

Fig 3. Maps of Plasmodium knowlesi malaria risk, human malaria elimination status, and model extrapolation versus interpolation.

A. Predicted risk of P. knowlesi malaria ranging from low to high risk. B. Countries projected to be malaria-free, eliminating malaria, or controlling malaria by 2025 (Map sourced from the University of California San Francisco Global Health Group’s Malaria Elimination Initiative) C. Comparison of environments in Malaysia, Brunei and Singapore (the model training region) with those across the rest of Southeast Asia, using all covariates and the multivariate environmental similarity surface (MESS) methods. The map distinguishes between areas of model interpolation and areas where the model was required to extrapolate to novel environments.