Patients and trial design
We retrospectively analyzed clinical, parasitological, and molecular data from all patients enrolled into the DHA-PPQ (DHA-PPQ) and artesunate-amodiaquine (ASAQ) arms in Bougoula-Hameau (southern Mali) of the West African Network for Clinical Trials of Antimalarial Drugs (WANECAM) clinical trial (PACTR201105000286876). The clinical trial was performed from 2011–2018 as a randomized, multicenter, open-label phase 3b/4 clinical trial at seven centers in Burkina Faso, Guinea and Mali. The DHA-PPQ and artesunate-amodiaquine arms in Bougoula were active from 2012 to 2015. Global primary and secondary endpoints have been previously reported3. The study was approved by local ethics committees and participating institutions (Ethical Committee of the Faculte de Medecin et D’Odonto-Stomatology, ref. 2010_79_FMPOS as well as by Stockholm Regional Ethics Committee, ref. 2016/2286-32 and 2017/499-32). Patients of either sex, aged ≥ 6 months with uncomplicated, microscopically confirmed Plasmodium spp malaria (axillary temperature ≥ 37.5 °C, or oral, or rectal, or tympanic temperature ≥ 38 °C or history of fever in the previous 24 h, and parasite density > 0 to < 200,000 parasites per μL blood) were randomly assigned to receive either DHA-PPQ (Alfasigma, Italy) or artesunate-amodiaquine (Sanofi, France) for each successive malaria episode during the 2-year follow-up of each patient (except for rescue treatment for episodes within 26 days after start of treatment) (Supplementary Fig 1). Exclusion criteria included a hemoglobin concentration of < 7 g/dL, non-malarial febrile conditions, anti-malarial treatment within the previous two weeks, significant liver or renal impairment, and pregnancy. Once daily dosing for 3 days was per bodyweight (Supplementary Table 2). Patients were hospitalized during the directly observed treatment phase. After inclusion and after at each subsequent episode patients were followed up actively (weekly by microscopy) and after 63 days, passively by inviting study participants to present for any subsequent fever episode for a period of two years. As originally reported in the WANECAM clinical trial3, recrudescence was distinguished from reinfection using msp1, msp2, and microsatellite markers. Briefly, for uncomplicated P. falciparum malaria, PCR-adjusted adequate clinical and parasitological response (ACPR) was greater than 99.5% at day 28 and greater than 98.6% at day 42 for all ACTs included.
Molecular analysis of plasmepsin genes pfpm2 and pfpm3
Genetic analysis was performed on samples collected at enrollment and upon parasite recurrence prior to re-treatment. DNA from dried blood spots on Whatman 3-mm filter paper (FP) was extracted with the QIAamp DNA Mini kit (Qiagen, Germany) according to the manufacturer´s instructions.
Pfpm2 (PF3D7_1408000) gene copy number was assessed by a SYBR green-dye-based real-time polymerase-chain-reaction (PCR) assay in triplicates using a modified version of a previously described protocol8,32 (Supplementary Table 3). Copy numbers of the pfpm3 (PF3D7_1408100) gene were assessed by a TaqMan probe-based protocol adapted from Ansbro et al. (2020)17 (Supplementary Table 3). The presence of the structural breakpoint of the plasmepsin 1/3 amplification, previously identified in SE Asia parasites, was assessed in every infection showing evidence of carrying plasmepsin 2 multiple copies7,17. The P. falciparum 3D7 strain was used as an external single gene copy control, the β-tubulin (PF3D7_1008700) gene served as an internal control, and a previously designed genetically modified clone with 2 copies of the pfpm2 and pfpm3 genes functioned as the control for multicopies32.
PCR-Restriction Fragment Length Polymorphism (PCR-RFLP) and Sanger sequencing methods were used for the analysis of pfcrt SNPs (K76T, H97Y, C101F, F145I, N326S, and I356T), previously associated with increased piperaquine tolerability (Supplementary Table 4).
Data analysis
The endpoints of our analyses were (i) the DHA-PPQ versus ASAQ relative risk of re-infection following antimalarial treatment in different study years, (ii) the time between treatment and re-infection, and (iii) the pfpm3 gene CNV of the infecting parasite in patients allocated to DHA-PPQ. In addition, pfpm3 gene copy number variation of artesunate-amodiaquine treatment arm infections was assessed as a comparison control arm.
Incidence density, or daily re-infection rates, were estimated using the total number of re-infection events over a 7-day period and normalized by the number of individuals in the respective trial arm, thus taking into account individuals entering and leaving the different treatment arms over time; data smoothing for visualization purposes was performed using R’s local polynomial regression (LOESS) function loess.
Risk ratios, estimated using unconditional maximum likelihood estimations as encoded in the riskratio() function of the epitools R library (https://doi.org/10.32614/CRAN.package.epitools), were determined by comparing 6-months cumulative re-infection rates of DHA-PPQ treated individuals against those treated with ASAQ. For this analysis we only considered the period 2012–2014 due to the lack of samples from the DHA-piperaquine arm for 2015.
Due to the strong seasonality in malaria transmission in these settings, only infections that occurred within the same year between March and December were included in the calculation of the time to reinfection. Statistical significance between year-stratified mean reinfection times or mean pfpm3 CNV was determined using Welch Two Sample t-tests or ANOVA, depending on the number of categories compared.
As the two study arms were analyzed with different batches of Taqman® probes for determining pfpm3 CNV, and because the focus of this study was on their relative changes over time, and in comparison between the two study arms, all CNV scores were normalized against the respective average scores based on the first 50 infections in each arm, to permit direct comparison of temporal trends in this mutation over the study duration.
Re-infection time distributions are presented using box and whisker plots, where boxes indicate the median (bar), first and third quartiles (hinges), and the lower and upper whiskers indicate the smallest and largest value no further than 1.5 times the interquartile range from the hinges. All data analysis was performed using the R open-source statistical software, version 4.2 (R Project for Statistical Computing).
Reporting summary
Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.
