Defining the In Vivo Phenotype of Artemisinin-Resistant Falciparum Malaria: A Modelling Approach: e1001823

• Published in: PLoS medicine 2015-04, Vol.12 (4)
• Main Authors: White, Lisa J, Flegg, Jennifer A, Phyo, Aung Pyae, Wiladpai-ngern, Ja Hser, Bethell, Delia, Plowe, Christopher, Anderson, Tim, Nkhoma, Standwell, Nair, Shalini, Tripura, Rupam, Stepniewska, Kasia, Pan-Ngum, Wirichada, Silamut, Kamolrat, Cooper, Ben S, Lubell, Yoel, Ashley, Elizabeth A, Nguon, Chea, Nosten, François, White, Nicholas J, Dondorp, Arjen M
• Format: Article
• Language: English
• Subjects: Confidence intervals; Genotype & phenotype; Infections; Malaria; Mathematical models; Parasites; Population; Studies; Surveillance
• ISSN: 1549-1277; 1549-1676
• DOI: 10.1371/journal.pmed.1001823

Background Artemisinin-resistant falciparum malaria has emerged in Southeast Asia, posing a major threat to malaria control. It is characterised by delayed asexual-stage parasite clearance, which is the reference comparator for the molecular marker 'Kelch 13' and in vitro sensitivity tests. However, current cut-off values denoting slow clearance based on the proportion of individuals remaining parasitaemic on the third day of treatment ('day-3'), or on peripheral blood parasite half-life, are not well supported. We here explore the parasite clearance distributions in an area of artemisinin resistance with the aim refining the in vivo phenotypic definitions. Methods and Findings Data from 1,518 patients on the Thai-Myanmar and Thai-Cambodian borders with parasite half-life assessments after artesunate treatment were analysed. Half-lives followed a bimodal distribution. A statistical approach was developed to infer the characteristics of the component distributions and their relative contribution to the composite mixture. A model representing two parasite subpopulations with geometric mean (IQR) parasite half-lives of 3.0 (2.4-3.9) hours and 6.50 (5.7-7.4) hours was consistent with the data. For individual patients, the parasite half-life provided a predicted likelihood of an artemisinin-resistant infection which depends on the population prevalence of resistance in that area. Consequently, a half-life where the probability is 0.5 varied between 3.5 and 5.5 hours. Using this model, the current 'day-3' cut-off value of 10% predicts the potential presence of artemisinin-resistant infections in most but not all scenarios. These findings are relevant to the low-transmission setting of Southeast Asia. Generalisation to a high transmission setting as in regions of Sub-Saharan Africa will need additional evaluation. Conclusions Characterisation of overlapping distributions of parasite half-lives provides quantitative insight into the relationship between parasite clearance and artemisinin resistance, as well as the predictive value of the 10% cut-off in 'day-3' parasitaemia. The findings are important for the interpretation of in vitro sensitivity tests and molecular markers for artemisinin resistance and for contextualising the 'day 3' threshold to account for initial parasitaemia and sample size.