Decreased dihydroartemisinin-piperaquine protection against recurrent malaria associated with Plasmodium falciparum plasmepsin 3 copy number variation in Africa

Abstract

Dihydroartemisinin-piperaquine (DHA-PPQ) is being recommended in Africa for the management of uncomplicated Plasmodium falciparum malaria and for chemoprevention strategies, based on the ability of piperaquine to delay re-infections. Although therapeutic resistance to piperaquine has been linked to increased copy number in plasmepsin-coding parasite genes (pfpm), their effect on the duration of the post-treatment prophylactic period remains unclear. Here, we retrospectively analyzed data from a randomized clinical trial, where patients received either DHA-PPQ or artesunate-amodiaquine for recurrent malaria episodes over two years. We observed an increase in the relative risk of re-infection among patients receiving DHA-PPQ compared to artesunate-amodiaquine after the first malaria season. This was driven by shorter average times to reinfection and coincided with an increased frequency of infections comprising pfpm3 multi-copy parasites. The decline in post-treatment protection of DHA-PPQ upon repeated use in a high transmission setting raises concerns for its wider use for chemopreventive strategies in Africa.

Fig. 1. Decline of long-term protection conferred by piperaquine.

A Estimated incidence density (number of infections / person / day) of clinical malaria episodes in the artesunate-amodiaquine (ASAQ, blue) and DHA-piperaquine (yellow) trial arms showing pronounced seasonality and decreasing transmission intensity in the study area together with a reduction in the long-term protective capacity of DHA-piperaquine, leading to converging incidence densities between the two arms over time. B Estimated risk ratios (DHA-piperaquine arm relative to the ASAQ arm, red circles) together with 95% confidence intervals demonstrating the decrease in long-term protection against reinfection over time with no statistically significant difference in (reinfection) risk from 06/2013 onwards. C Boxplot showing the decrease in the average time intervals between treatment and subsequent infection between 2012 and 2014 in the DHA-piperaquine arm (P < 0.002, ANOVA), but not the ASAQ control arm (P = 0.9, ANOVA).

Fig. 2. Selection of multi-copy pfpm3 parasites in the DHA-PPQ arm.

A Bar chart showing the number of clinical episodes in the DHA-PPQ trial arms, color-coded by average pfpm3-CNV, illustrating how selection appears to intensify along the high-transmission season (May-December) and towards the end of the trial, likely in response to an increase in drug pressure. B Boxplot of normalized pfpm3 CNV demonstrating an increase of multi-copy variant infections between 2012 and 2014 in the DHA-piperaquine (P < 0.001, ANOVA); the overall trend in the ASAQ arm is slightly negative, driven by a temporary decline in 2013 (P = 0.001, ANOVA).

Fig. 3. pfpm3 gene dose effect vs time to reinfection following DHA-piperaquine and ASAQ treatment.

Increased average pfpm3 copy number is linked with a decrease in the time interval between treatment and reinfection, with early re-infections enriched with parasites carrying more than one pfpm3 gene copy in the DHA-piperaquine arm (orange dots); this effect is not found in the ASAQ arm (blue dots). CNVs are scaled to aid comparison between both arms. Local regression lines (LOESS) are added to guide the eye; shaded areas indicate 95% confidence intervals.