Extending seasonal malaria chemoprevention to older children: a non-inferiority randomized controlled trial of Dihydroartemisinin–piperaquine as a policy alternative in Mali

Summary

Background

Seasonal malaria chemoprevention (SMC) with Sulfadoxine-pyrimethamine plus amodiaquine (SPAQ) reduces malaria in children under five, but substantial burden remains among children 5–9 years. Emerging SP/AQ resistance and adherence challenges underscore the need for alternative regimens. Dihydroartemisinin-piperaquine (DHAPQ) is a promising option, although evidence from multi-season programmatic settings in the Sahel is limited. We aimed to evaluate the effectiveness and safety of DHAPQ compared with SPAQ when SMC was extended to children aged 3 months to 9 years.

Methods

We conducted a two-year, three-arm, cluster-randomized crossover trial in nine health-catchment clusters in Mali. Clusters received standard SPAQ SMC (<5 years), extended-age SPAQ (<10 years), or (3) extended-age DHAPQ (<10 years), crossing over to a different regimen in year two. The primary outcome was the incidence of clinical malaria, analyzed using mixed-effects negative binomial models that accounted for clustering, period effects, and recurrent events. Non-inferiority was predefined as an upper 95% CI < 1.25.

Findings

Among 12,854 children, DHAPQ was non-inferior to extended-age SPAQ (HR = 0.7, 95% CI 0.6–0.9) and achieved higher intervention coverage, with no serious adverse events reported. Children aged 5–9 years had a lower hazard of clinical malaria than children younger than 5 years (HR = 0.9, 95% CI 0.8–0.9).

Interpretation

Extending SMC to children aged 5–9 years substantially reduced malaria burden. DHAPQ was safe, effective, and operationally feasible, supporting its consideration as an alternative to SPAQ where adherence or resistance threatens SMC effectiveness.

Funding

NIH/NIAID—ICEMR-WAF.

Introduction

Global progress against malaria has stalled since 2015, with cases increasing by 22.6%, nearly 80% of this rise occurring in the WHO African regions.¹ Population growth, conflict, climate variability, and improved surveillance have contributed to this trend.^1,2 To reduce malaria burden among vulnerable populations, the World Health Organization (WHO) recommends preventive strategies including Intermittent preventive treatment in pregnancy with Sulfadoxine-pyrimethamine (IPTp-SP) and Seasonal malaria chemoprevention (SMC) with Sulfadoxine-pyrimethamine plus amodiaquine (SPAQ) for children under 5 years living in areas of highly seasonal transmission.³ Since its recommendation in 2012, SMC has reduced uncomplicated and severe malaria by approximately 75% during the transmission season.^4,5

In Mali, malaria transmission season peaks between July and October.^6,7 SMC with SPAQ has been implemented since 2016 through four monthly treatment rounds for children under five.⁸ Although this strategy has substantially reduced malaria morbidity among children under five,^8,9 an increasing burden of asymptomatic infection and clinical malaria has been documented among children aged 5–9 years, who are not currently eligible for SMC.¹⁰

The long-term effectiveness of SPAQ-based SMC is a; so threatened by emerging resistance to SP and AQ, poor adherence to the three-day AQ regimen, adverse drug reactions, and limited pharmacovigilance. Although resistance markers have not compromised SPAQ effectiveness in Mali,¹¹ emphasizing the need to evaluate alternative regimens and expanded age eligibility. Dihydroartemisinin-piperaquine (DHAP) has demonstrated high efficacy, good tolerability and prolonged post-treatment prophylaxis, and a trial in Burkina Faso showed it was as effective as SPAQ in preventing malaria among children under five.¹² However, evidence from high-transmission Sahelian settings remains limited, particularly regarding multi-season effectiveness and use in older children. We therefore conducted a two-year cluster-randomized trial in Koulikoro, Mali, to evaluate DHAPQ as an alternative to SPAQ while extending SMC to children aged 5–9 years old. Our findings provide policy-relevant evidence to inform future SMC strategies in areas of high seasonal malaria transmission.

Methods

Study sites

The study was conducted in the Koulikoro health district, Mali, a Sudan-Sahelian area with highly seasonal malaria transmission and annual rainfall of 600–1200 mm (Fig. 1). The district comprises 21 community health centers (CSCOM) serving approximately 282,570 people. Malaria transmission varies across three ecological zones: high and prolonged transmission along the river Niger (5–6 months); moderate transmission in the central zone lasting 3–4 months; and lower transmission in the northern zone (2–3 months). Anopheles gambiae sensu lato (s.l.) is the primary vector. Routine malaria control includes insecticide-treated nets (ITNs), prompt diagnosis and treatment, intermittent preventive therapy in pregnancy (IPTp), SMC for children under 5, and R21/Matrix-M vaccination for children under three.

Fig. 1:

Cross over randomization Map in 2019 and 2020.

Study design

A three-arm, cluster-randomized crossover trial was conducted over two consecutive malaria transmission seasons (2019–2020) in nine community health zones in Koulikoro, Mali. Health zones were selected using stratified random sampling with probability proportional to population size¹³ to ensure representation across three (3) ecological strata (riverine, central, and northern zones). The 9 zones were: Gouni, Kenenkou and Kamani (Riverine), Doumba, Koula and Sinzani (Central), Sirakorola, Monzombala and Chola (northern).

Eligibility and enrolment

All children aged 3 months to 9 years without chronic illness or known drug allergy were eligible. Baseline data were collected at enrolment each year, prior to the start of the malaria transmission season (June) and before any exposure to SMC. Enrolment was conducted including all eligible children within each cluster following parental consent. Children aged out of eligibility were excluded in the subsequent year, and newly eligible children were enrolled. These assessments were conducted at the community health center (CHC) in each village rather than during home visits.

Study comparison groups

The trial used a three-arm, cluster-randomized design in which nine clusters were allocated to one of three SMC delivery strategies. The comparison groups were defined as follows:

• Arm 1: Control (Standard SMC), children <5 years received SPAQ through the national SMC programme, while children aged 5–9 years did not received SMC. This arm represents the current standard of care and serves as the reference for evaluating extended-age SMC strategies.

• Arm 2: SPAQ <10 (Extended-age SPAQ): in this arm, the same SPAQ regimen was delivered to children aged 3 months to 9 years. The only difference from Arm 1 is age eligibility, not drug regimen or delivery strategy.

• Arm 3: DHAPQ (Extended-age DHAPQ): Children aged 3 months to 9 years received DHAPQ instead of SPAQ

All arms followed the same delivery procedures: day-1 doses administered by community health workers under direct observation, with days 2–3 administered by caregivers.

To strengthen internal validity and reduce confounding by stable ecological differences, a predefined crossover was implemented in year 2. Each cluster was reassigned to a different intervention according to a fixed rotation scheme, allowing every cluster to contribute data under more than one regimen (see Fig. 1).

The primary comparison for the non-inferiority analysis was DHAPQ vs SPAQ <10 years old, as both regimens targeted the same age group and differed only in drug choice. Comparisons with the Control arm were conducted using cluster-adjusted mixed-effects models to evaluate the added value of extending SMC beyond the current age range.

Secondary comparisons included:

1. Arm 2 vs Arm 1 to evaluate the effect of extending SMC to children aged 5–9 years; and

2. Arm 3 vs Arm 1 to assess the combined effect of age extension and alternative drug regimen.

A non-inferiority design was selected because DHAPQ is already known to be highly efficacious, and the key policy question was whether DHAPQ could provide at least comparable protection to SPAQ while offering potential advantages in adherence, tolerability, and operational feasibility. A superiority design was not appropriate because: (1) DHAPQ’s known efficacy does not guarantee superiority under programmatic, multi-season conditions; (2) DHAPQ carries potential risks, including tolerability concerns and the possibility of accelerating artemisinin resistance, that require demonstrating no clinically meaningful loss of protection before considering policy change; and (3) non-inferiority is the recommended framework when evaluating alternative regimens intended to replace or complement an established standard of care.

A superiority design was not adopted because the key policy question was whether DHAPQ provides no clinically meaningful loss of protection compared with SPAQ, while offering advantages in adherence and tolerability, and without increasing resistance risks, making non-inferiority the appropriate methodological framework. The trial was designed and reported in accordance with the CONSORT guidelines for cluster randomized trials (CONSORT checklist, Supplementary File 2).

Intervention procedures

In all study arms, SMC delivery followed the national policy with four monthly rounds implemented from July to October. During each round, field teams were deployed at each study site to supervise drug distribution and monitor adherence. Community health workers and community health agents were trained before the first round on SMC drug administration, dosing schedules, use of rapid diagnostic tests, and electronic data capture procedures. Children received SPAQ (Arm 1 and 2) or DHAPQ (Arm 3). Day-1 doses were administered at the household level by community health agents under directly observed therapy (DOT), with project staff present for quality assurance. Days 2 and 3 doses were administered at home by caregivers. This procedure was identical across all arms, including the Control arm. Both drug dosing was age-based, consistent with Mali NMCP guidelines. As per WHO recommendations, children were observed for 15 min after the first dose to monitor for vomiting or immediate adverse reaction to the treatment; doses were re-administered if vomiting occurred within this period. Dosing was provided according to WHO recommendation as follows: for SPAQ: 25 mg/kg Sulfadoxine and 1.25 mg/kg Pyrimethamine, plus amodiaquine 10 mg/kg daily for three days. For DHAPQ: 4 mg/kg DHA and 18 mg/kg PQ daily for three days. Children presenting symptomatic malaria at the time of SMC administration were tested using malaria rapid diagnostic test (mRDT) and treated with Artemether Lumefantrine (ALu), per national guidelines.

Outcomes assessment

The primary outcome was the incidence rate of clinical malaria episodes during follow-up. The secondary outcomes included SMC coverage, adverse events, asymptomatic parasitemia prevalence, anemia, and entomological outcomes, including mosquito density, biting rates, longevity, population age structure, sporozoite rate, and entomological inoculation rate (EIR).

Incidence of Clinical malaria

Malaria incidence was measured through passive case detection at the community health centers. Caregivers were encouraged to bring children for any illness, particularly fever or any malaria-related symptoms. A trained clinical investigator was placed at each health center to perform standardized clinical examination, malaria RDT, blood smear for microscopy, and dried blood spots (DBS) during each visit. Clinical malaria was defined as axillary temperature ≥37.5 °C plus laboratory-confirmed parasitemia (mRDT and/or microscopy). All confirmed malaria cases were treated free of charge with respect to national guidelines for malaria treatment.

SMC coverage

In line with the Mali NMCP guideline, all children, regardless of the study arm, received only the first dose of either SPAQ or DHAPQ under Directly observed therapy (DOT) by a community health worker supervised by the research team. Coverage was defined in this study as the proportion of children who received the first dose of SMC treatment per month and per cluster, consistent with NMCP programmatic indicators.

Adverse events

Adverse events and treatment adherence were assessed through post-SMC household surveys conducted after each treatment round among randomly selected participating households. Caregivers were interviewed regarding adherence to the treatment regimen, adverse side effects, and reasons for non-compliance with SMC treatment.

Reported symptoms included nausea, vomiting, headache, malaise, weakness, diarrhea, chills, cough, and other caregiver-reported symptoms.

Any suspected severe adverse event was referred to the nearest health facility and managed according to national malaria treatment guidelines. These procedures were specified in the study protocol registered at ClinicalTrials.gov (NCT04149106).

Parasitemia assessments

Cross-sectional parasitemia assessments were conducted before each SMC cycle in both years, in June (baseline) and before each SMC round (July, August, September, and October). A random sample of enrolled children using proportional-to-population-size sampling was tested by microscopy and hemoglobin measurement, irrespective of recent illness or malaria history. Prevalence of asymptomatic parasitemia was defined as the proportion of children with asexual and/or sexual Plasmodium falciparum detected by microscopy only. Anemia was also measured during each cross-sectional survey using a Hemocue 301^® and defined as hemoglobin concentration (Hb) < 11 g/dl in this study.

Entomological surveys

Adult mosquitoes were collected by pyrethrum spray catch (PSC) from 360 randomly (120 per study arm) selected concessions (groups of households sharing the same meal). Six surveys were carried out in each village over the 2 year-period (2019–2020). Mosquitoes were identified morphologically, sorted by abdominal status (unfed, fed, half-gravid, and gravid), and dissected into two portions (head-thorax and abdomen), dissected and stored in ethanol. Blood meal sources were identified by a direct enzyme-linked immunosorbent assay (ELISA) using human antibodies. Head and thorax portions of the preserved Anopheles specimens were tested for P. falciparum circumsporozoite proteins (CSPs) using the ELISA. Standard indicators, including vector density, human biting rate (HBR), infection rate (IR), human blood index (HBI), and entomological EIR, were calculated.

Statistical power and sample size

The sample size for the primary non-inferiority comparison (Arm 3 [DHAPQ] vs. Arm 2 [SPAQ]) was calculated using the Schoenfeld formula for time-to-event end-points.¹⁴ Specifically, $D=\frac{4{\left(z_{1-\alpha /2}+z_{1-\beta }\right)}^2}{{\left[ln\left(H{R}_{NI}\right)-ln\left(H{R}_{\mathit{Epected}}\right)\right]}^2}$, where where $D$ denotes the number of required events, $z_{1-\alpha /2}$ and $z_{1-\beta }$ represent the critical values corresponding to the type I and type II error rates, and $H{R}_{NI}$ and $H{R}_{\mathit{Epected}}$ denote the noninferiority and expected hazard ratios, respectively. We applied a two-sided type I error threshold of 5% (i.e., $z_{1-\alpha /2}=z_{0.975}=1.960$), which is equivalent to a one-sided threshold of 2.5% and represents a conservative approach compared with the commonly used one-sided 5% convention for non-inferiority trials. With a type II error threshold of 20% (power = 80%, $z_{1-\beta }=0.842$), an allocation ratio of 50%, a non-inferiority margin of 1.25, and an expected hazard ratio of 1, the total number of participants required for the two-arm is 631/0.50 = 1262 participants (631 per arm). The sample size was powered for the primary non-inferiority comparison between Arm 3 (DHAPQ) and Arm 2 (SPAQ), which constitutes the key study hypothesis. The third arm (Control, Arm 1) was included to support pre-specified secondary superiority comparisons; no separate sample size calculation was performed for these comparisons, which were designated as exploratory, and no formal adjustment for multiple comparisons was applied.

To account for the cluster-randomized design, we applied the standard design effect inflation formula. Based on an anticipated total enrolment of 6000 children distributed across 9 equal clusters, the expected average cluster size was m = 6000/9 ≈ 667 participants per cluster. The intraclass correlation coefficient (ICC, ρ) was estimated at 0.01, based on prior malaria cohort data from the same district. The design effect was: DEFF = 1+(m−1) × ρ = 1+(667−1) × 0.01 = 7.66 rounded up to 8 to provide a conservative estimate and preserve the target power of 80%. Applying this design effect to the per-arm requirement yields 632 × 8 = 5056 participants per arm (cluster-adjusted), and 5056 × 2 = 10,112 participants for the two-arm non-inferiority comparison. For operational planning purposes across the three study arms, the total cluster-adjusted requirement is 5056 × 3 = 15,168 participants. A total of 12,854 children were enrolled across the three arms (approximately 4285 per arm), providing approximately 87% of the cluster-adjusted target for the primary two-arm non-inferiority comparison.

Statistical analysis

The primary analysis followed an intention-to-treat (ITT) principle, whereby all participants were analyzed according to the intervention assigned to their cluster, regardless of adherence. In addition, per-protocol (PP) analyses were conducted for the non-inferiority comparison between DHAPQ and SPAQ to evaluate the consistency and robustness of the findings. The PP population included children who received the assigned intervention, completed the scheduled SMC regimen, had primary outcome data available, and had no major protocol deviations. Major protocol deviations included receipt of a non-assigned regimen, incomplete treatment administration, or missing primary outcome follow-up data.

Baseline characteristics were summarized descriptively by study arm. Categorical variables were presented as frequencies and percentages without formal hypothesis testing, as randomization occurred at the cluster level. SMC coverage was summarized descriptively by study arm, month, year, and age group using proportions and percentages. Adverse events were summarized descriptively by study arm and treatment round. Binary logistic regression models were used to estimate odds ratios and 95% confidence intervals for comparisons between study arms; these analyses were considered exploratory.

The primary outcome was the incidence rate of clinical malaria episodes, analyzed using mixed-effects negative binomial regression models to account for overdispersion and repeated malaria episodes during follow-up. Study arm, study period (year), treatment-by-period interaction, and prespecified covariates were included as fixed effects. Random intercepts for participant and village (cluster) were included to account for repeated observations within participants and intra-cluster correlation, respectively. Log person-time was included as an offset term to model incidence rates. Results are presented as incidence rate ratios (IRRs) with 95% confidence intervals (CI). Kaplan–Meier survival curves were generated as secondary descriptive analyses to illustrate malaria-free survival over the 30-day period following SMC administration according to study arm.

For the non-inferiority time-to-event analysis, hazard ratios (HR) and 95% confidence intervals were estimated using Cox proportional hazards regression models. Time-to-event was defined as the number of days from enrolment to the first episode of clinical malaria. Participants who did not experience a clinical malaria episode were censored at the end of their observed follow-up period. The models included study arms, age groups, gender, and study year as covariates. To account for the cluster-randomized design, village-level cluster-robust standard errors were used.

Parasite prevalence was analyzed using mixed-effects logistic regression models, with study arm, study period (year), treatment-by-period interaction, and prespecified covariates included as fixed effects, and village included as a random intercept to account for intra-cluster correlation. Results are presented as odds ratios (ORs) with 95% CI. Hemoglobin concentrations were summarized using descriptive statistics and modeled using linear mixed-effects models with fixed effects for age, calendar month, and study arm, and a random intercept for cluster to account for intra-cluster correlation.

Adverse events were summarized as the number of children reporting at least one event within 5 days of SMC administration.

A non-inferiority framework was used to compare the two extended SMC regimens (DHAPQ vs. SPAQ <10), with SPAQ serving as the reference group. The non-inferiority margin was prespecified as a HR of 1.25, to preserve a substantial proportion of the established protective effect of SPAQ while allowing for a limited and clinically acceptable increase in malaria risk under routine programmatic conditions. Non-inferiority conclusions were assessed for consistency across both analysis populations, in accordance with methodological recommendations for non-inferiority trials. All analyses accounted for the cluster-randomized design through inclusion of cluster-level random effects.

Data were collected electronically using REDCap with built-in validation rules and consistency checks to ensure data quality. Data were analyzed in R version 4.4.3. The type I error threshold was set at 5%.

Entomological indicators were computed following standard definitions. Vector density was expressed as the mean number of indoor resting mosquitoes per room per day; the human biting rate (HBR) was calculated as the average number of mosquito bites received by a sleeper per time unit (blood-fed and half gravid mosquito/number of room sleepers). The IR corresponds to the proportion of An. gambiae s.l. carrying P. falciparum sporozoites. The HBI was defined as the proportion of female mosquitoes having human blood in their guts. The EIR was calculated as the total number of infectious bites per person per unit time.

Missing data

Missing data arose mainly from incomplete follow-up for the primary outcome and from non-observed doses on days 2 and 3 of each SMC treatment round. For the primary analysis, we used an intention-to-treat approach in which all enrolled participants were analysed according to the intervention assigned to their cluster, and individuals contributed person-time only for the period they were under observation; no statistical imputation of missing outcomes was performed. The per-protocol population was restricted to children who received the assigned regimen, completed the scheduled SMC course, had primary outcome data available, and had no major protocol deviations, including missing primary outcome follow-up data. For baseline characteristics and secondary outcomes, categorical variables were summarised using non-missing observations for each variable, so denominators varied according to the amount of observed data.

Ethical considerations

This study was registered on clinicaltrial.gov under the number NCT04149106, and the study protocol was approved by the institutional ethics committee of the Malian University of Sciences, Techniques and Technologies of Bamako (USTT-B) under the reference number N*19/7CE/FMOS/FAPH. Written informed consent was obtained from the parents and/or legal guardians of the participants. Study conduct and participant safety were monitored by the study coordination team throughout the trial. An independent auditor supervised the implementation of the study procedures and data quality. All adverse events were recorded and managed according to national pharmacovigilance guidelines. The trial was registered retrospectively due to administrative shorter time between Notice of Awards plus funds availability and the beginning of the first year SMC campaign in 2019. However, no changes were made to the study design, outcomes, or statistical analysis plan at the time of registration. The administrative delay in registration did not result in any protocol deviations, nor did it affect participant enrollment, intervention delivery, or outcome measurement. This trial was conducted in accordance with the principles of the Declaration of Helsinki, the International Council for Harmonization Good Clinical Practice (ICH-GCP) guidelines, and all applicable national regulatory requirements.

Results

Study population and baseline characteristics

A total of nine clusters were randomized and allocated to the study arms, with three clusters allocated to each study arm. All clusters were included in the analysis. Crossing these clusters, 12,976 children were assessed for eligibility, of whom 12,854 met the inclusion criteria and were enrolled. The number of children assessed and enrolled within each arm reflected the population size of the corresponding clusters (Table 1).

Table 1:

Baseline demographic and clinical characteristics of participants by study arm and year before the first SMC round each year.

Characteristics	2019				2020
	Control	SPAQ	DHAPQ	Overall	Control	SPAQ	DHAPQ	Overall
Sample size, n	1946	2189	2068	6203	2087	2415	2149	6651
Gender n (%)
Male	958 (49.2)	1097 (50.1)	1091 (52.8)	3146 (50.7)	1078 (51.7)	1219 (50.5)	1088 (50.6)	3385 (50.9)
Female	988 (50.8)	1092 (49.9)	977 (47.2)	3057 (49.3)	1009 (48.3)	1196 (49.5)	1061 (49.4)	3266 (49.1)
Fever n (%)	98 (5.2)	127 (5.9)	117 (5.8)	342 (5.6)	382 (18.3)	534 (22.1)	176 (8.2)	1092 (16.5)
Anemia n (%)	311 (16.4)	609 (28.2)	395 (21.5)	1315 (22.3)	348 (18.0)	337 (17.8)	464 (23.2)	1149 (19.7)
P. falciparum infection n (%)	130 (7.4)	294 (13.6)	197 (11.0)	621 (10.9)	247 (15.7)	161 (13.1)	248 (14.8)	656 (14.6)

Control: Standard SMC with SPAQ in children aged <5 years; SPAQ: Extended SMC with SPAQ in children aged ≤9 years; DHAPQ: Extended SMC with DHAPQ in children aged ≤9 years. Percentages were calculated using non-missing observations for each variable; denominators may therefore vary across characteristics.

Baseline characteristics were broadly comparable across study arms (Table 1). The proportion of males ranged from 49.2% (Control arm) to 52.8% (DHAPQ arm) in 2019, and a similar pattern was observed in 2020 (varying from 51.7 to 50.6%). Fever prevalence was low and comparable across groups in 2019 (5.2–5.9%) and higher with greater variability across arms in 2020 (8.2% in the control arm, reaching 22.1% in the SPAQ arm). The prevalence of anemia varied across arms in both years (2019: 16.4%–28.2%; 2020: 17.8%–23.2%). The prevalence of asymptomatic parasitemia also varied across study arms in 2019 (7.4%–13.6%) and was higher but more similar between arms in 2020 (13.1%–15.7%).

The flow of clusters and participants through the trial is presented in the CONSORT flow diagram (Fig. 2). Overall, 6203 children were enrolled in 2019 and 6651 in 2020, indicating balanced enrollment across the two study years. In the control arm, 2181 children aged <5 years received standard SMC with SPAQ, while 1869 children aged 5–9 years were not eligible for SMC under the national policy. In the SPAQ arm, 4604 children aged 3 months to 9 years received monthly SMC with SPAQ. In the DHAPQ arm, 4217 children aged 3 months to 9 years received monthly SMC with DHAPQ. The allocation of clusters across study arms differed between 2019 and 2020 as per the crossover design.

Fig. 2:

CONSORT flow diagram of clusters and participants through the trial.

SMC coverage

The SMC coverage rate, defined as the proportion of eligible children who received SMC on the first day under DOT each month, was consistently high (Supplementary Figure S1). In 2019, SMC coverage among children <5 years ranged from approximately 72%–88% from July to October, with consistently higher uptake in the DHAPQ arm (p < 0.008 regardless of the month). In children aged 5–9 years (intervention arms only), coverage increased from August to October from 74% to nearly 88%, again highest in the DHAPQ arm (p < 0.007 regardless of the month).

In 2020, a substantial improvement in coverage was observed across study arms. Among children <5 years, coverage exceeded 85% as early as August and reached up to 91.8% in October in the DHAPQ arm. In the extended-age group (5–9 years), coverage also increased compared with 2019, with the DHAPQ arm surpassing 90% by October. Overall, the DHAPQ arm achieved the highest and most stable coverage levels across both years and age groups (Supplementary Figure S1).

Adverse events

The forest plot Supplementary Figure S3 (Supplementary File) compares the frequency of key adverse events between the control group and children receiving SPAQ or DHAPQ. Both treatment groups report common symptoms (notably nausea-vomiting, headache, malaise, and weakness) more often than the control arm. Overall, SPAQ appears to be associated with a higher reporting of several side effects compared with the control group. DHAPQ shows a broadly similar side-effect profile, but with slightly lower or comparable frequencies than SPAQ for most symptoms.

No SAEs, including severe anemia and hospitalization, were reported during the two years of trial.

Clinical malaria incidence

In the primary mixed-effects negative binomial regression analysis (Table 2), children receiving DHAPQ experienced a significantly lower incidence of repeated clinical malaria episodes compared with the control arm (IRR = 0.8; 95% CI 0.8–0.9; p = 0.002). In contrast, no significant difference was observed between the SPAQ arm and the control arm (IRR = 1.0; 95% CI 0.9–1.1; p = 0.948).

Table 2:

Incidence of repeated clinical malaria episodes by study arm: Mixed-effects negative binomial regression analysis.

Variable	Category	IRR (95% CI)	p-value
Study arm	Control	Reference
	SPAQ	1.0 (0.9–1.1)	0.948
	DHAPQ	0.8 (0.8–0.9)	0.002
Year	2019	Reference
	2020	0.9 (0.8–1.0)	0.059
Age group	3–59 months	Reference
	5–9 years	1.0 (0.9–1.10)	0.360
Sex	Female	Reference
	Male	1.0 (0.9–1.1)	0.860

Across study years, malaria episode incidence did not differ significantly (IRR = 0.9; 95% CI 0.8–1.0; p = 0.059). Similarly, age group and sex were not significantly associated with repeated clinical malaria episodes.

Regardless of years and age groups, seasonal patterns of clinical malaria were evident across all arms (Fig. 3). In 2019, however, a significant variation in cumulative incidence was observed in the peak season (October), reaching 25% in the control arm compared to 22 and 7% in SPAQ and DHAPQ arms, respectively. A similar pattern was observed in children <5 years, with lower magnitudes. In 2020, overall incidence levels were lower, but the ranking persists: DHAPQ had the lowest monthly incidence throughout the season in both age groups, SPAQ showed intermediate values, and the Control arm exhibited the highest peaks, especially 20% in November among 5- to 9-year-olds.

Fig. 3:

Monthly malaria incidence by study arm, year, and age group.

Survival to the first malaria episode was higher in the DHAPQ arm compared with the SPAQ arm after 5 months of follow-up (Fig. 4).

Fig. 4:

Time to malaria infection during the 30-day SMC cycle by study arm from 2019 to 2020.

The Kaplan–Meier curves demonstrated higher malaria-free survival in DHAPQ regimen across both years, followed by SPAQ arm, while the Control arm, limited to SPAQ <5, exhibited the lowest survival probabilities during the 30-day interval between consecutive SMC rounds. In both years, clear and consistent differences were observed across study arms. Curve separation occurred early after each SMC round and widened over the 30-day interval. Differences between arms were statistically significant (log-rank p < 0.0001).

Assessment of non-inferiority between the three SMC treatment regimens

In the primary non-inferiority analysis, children receiving DHAPQ had a lower hazard of clinical malaria compared with those receiving SPAQ (HR = 0.7, 95% CI 0.6–0.9), with the upper bound of the confidence interval remaining below the predefined non-inferiority margin of 1.25 (Fig. 5). Children aged 5–9 years had a slightly lower hazard of clinical malaria compared with children aged <5 years (HR = 0.9, 95% CI 0.8–0.9). Male and female participants had similar hazards of clinical malaria (HR = 0.9, 95% CI 0.9–1.0). Compared with 2019, the hazard of clinical malaria was lower during 2020 (HR = 0.7, 95% CI 0.5–0.9).

Fig. 5:

Non-inferiority assessment of SPAQ and DHAPQ study arms.

Additional intention-to-treat comparisons across study arms, including comparisons relative to the control arm, are provided in Supplementary Figure S2.

Asymptomatic parasitemia prevalence

The prevalence of Asymptomatic parasitemia varied significantly across study arms, age groups, and years (Table 3). In 2019, among children under 5 years, prevalence was 8.7% in the Control arm and 12.5% in the SPAQ arm (OR = 2.7, 95% CI 0.7–9.8), and markedly lower in the DHAPQ arm (4.8%, OR = 0.6, 95% CI 0.1–2.2). In the 5–9-year age group, SPAQ and DHAPQ reduced the odds of infection from 2.2 to 0.6 relative to the control arm.

Table 3:

Prevalence of asymptomatic parasitemia by study arm.

Year	Age group	Study arm	Positive case (n)	Prevalence (%)	OR [95% CI]
2019	<5 yo	Control	69	8.7	Ref
		SPAQ <10	69	12.5	2.7 [0.7–9.8]
		DHAPQ <10	28	4.8	0.6 [0.1–2.1]
	5–9 yo	Control	35	12.6	Ref
		SPAQ <10	120	21.5	2.2 [0.9–5.4]
		DHAPQ <10	56	9.2	0.6 [0.2–1.6]
2020	<5 yo	Control	194	12.6	Ref
		SPAQ <10	101	8.4	0.5 [0.2–1.2]
		DHAPQ <10	151	10.8	0.6 [0.3–1.5]
	5–9 yo	Control	248	19.9	Ref
		SPAQ <10	115	8.8	0.3 [0.1–0.7]
		DHAPQ <10	210	14.4	0.6 [0.3–1.3]

In 2020, prevalence increased across all groups, reaching 12.6% in the Control arm among < 5-year-olds. SPAQ significantly reduced the prevalence to 8.4% (OR = 0.5, 95% CI 0.2–1.2), while DHAPQ showed 10.8% (OR = 0.6, 95% CI 0.3–1.5). Among children aged 5–9 years, the Control arm showed the highest prevalence (19.9%), whereas substantial reductions were observed in SPAQ (8.8%, OR = 0.3, 95% CI 0.1–0.7) and DHAPQ (14.4%, OR = 0.6, 95% CI 0.3–1.3).

Hemoglobin concentrations and anemia

The linear mixed-effects model showed that hemoglobin levels increased progressively with age and exhibited a clear seasonal decline during peak transmission months Supplementary Figure S4. A seasonal pattern was also observed, with hemoglobin levels declining in September and October, corresponding to the highest malaria transmission. Median hemoglobin concentrations remained between 10 and 12 g/dL throughout the season, consistently close to the WHO anemia threshold of 11 g/dL, confirming the high burden of anemia previously observed in the cohort. Across the three study arms, hemoglobin trajectories largely overlapped, indicating no major difference attributable to treatment regimen. A slight elevation of hemoglobin levels was noted in the SPAQ and DHAPQ arms during September and October compared with the Control arm.

Entomological indicators

Entomological indicators of malaria transmission assessed in 2020 revealed marked seasonal increases in An. gambiae s.l. densities peaking in September across all study arms. Vector density rose steadily from July to September, with the highest peaks observed in the DHAPQ arm. The HBI remained very high across both study arms, confirming strong anthropophagy, with a temporary decrease in the Control arm in September. Overall, human–vector contact remained intense throughout the study period in all three study arms.

Entomological inoculation rate (EIR), known as a key indicator for transmission intensity, were defined as the number of infectious bites an individual receives from mosquitoes per unit of time. EIR varies substantially by arms: In the control arm, it remained low and stable from July (0.09) to August (0.04), sharply increased in September (0.66) and October (0.72). In the SPAQ year’s arm, the EIR increased progressively in August (0.20) and September (0.28), reaching a peak in October (0.46). In the DHAPQ arm, it was consistently high, increasing from 0 in July to 0.365 in August, peaking markedly in September (2.47), and subsequently declining in October (1.40). Overall, DHAPQ arm health zone experienced substantially higher EIRs, reflecting the ecological differences rather than drug effects, and underscoring the intense transmission environment (Fig. 6).

Fig. 6:

Monthly variations of Entomological indicators by study arms in Koulikoro in year 2020.

Discussion

While SMC with SPAQ for children under five remains a cornerstone of malaria control strategy in the Sahel,^12,15 yet its long-term sustainability is increasingly threatened. Rising frequencies of P. falciparum resistance markers to SP and AQ in several parts of sub-Saharan Africa,^16,17 combined with its side effects and acceptability among caregivers, underscore the urgency of identifying alternative regimens. In addition, epidemiological shifts have revealed that children aged 5–14 years (who are not targeted by current SMC policies in most endemic countries) now represent a major reservoir of both asymptomatic parasitemia and symptomatic malaria.¹⁸ Expanding SMC to this age group and evaluating effective and longer-acting drugs are therefore critical priorities for sustaining malaria control gains.

Piperaquine (PQ) is an antimalarial drug that presents with high effectiveness and a long duration of action, making it an ideal candidate for SMC.¹⁹ This drug is now available in combination with Dihydroartemisinin, a combination of antimalarial drugs that is 98% effective in the curative treatment of malaria.²⁰ DHAPQ has been used for chemoprevention in The Gambia²¹ and evaluated in high-transmission African settings, showing strong protective effects.²² Building on this evidence, this study assessed the effectiveness, tolerability, and operational feasibility of DHAPQ for SMC in children up to 9 years of age in Mali.

Over both years, SMC coverage in the DHAPQ arm consistently exceeded 80% and increased steadily over the season. In contrast, coverage in the SPAQ arm declined after the first round, remaining below 80% in later months. This pattern likely reflects the superior adherence of caregivers to DHAPQ because of the low frequencies of side effects among children following administration, as well as the strawberry taste of the DHAPQ, which makes it much easier to ingest.²³ Indeed, nausea, vomiting, and headache were substantially more common in SPAQ recipients, consistent with previous studies in Mali and Burkina Faso.^24,25 The markedly lower frequency of adverse events in the DHAPQ arm reinforces its suitability for monthly prophylaxis.

Children receiving DHAPQ experienced significantly fewer repeated clinical malaria episodes compared with the control arm, whereas no difference was observed between SPAQ and the control arm. These findings suggest that DHAPQ may provide improved protection against recurrent malaria episodes under routine implementation conditions. The absence of a significant reduction in the SPAQ arm may reflect differences in adherence, tolerability, transmission intensity, or other operational factors not fully captured in the analysis. Given that both SPAQ delivery approaches were implemented by the same team, the higher incidence observed in 2019 but not reproduced in 2020 likely reflects inter-annual variation in transmission intensity or stochastic variation, rather than a true difference in intervention effectiveness. Although clinical follow-up procedures were identical across arms, children in the intervention arms received monthly household visits for drug administration, whereas children aged 5–9 years in the Control arm did not. This difference could introduce a modest risk of differential ascertainment of clinical episodes. However, all children had continuous access to free diagnosis and treatment at study-supported health centers, and cross-sectional surveys included all arms, which likely minimized this potential bias.

In the primary per-protocol non-inferiority analysis, DHAPQ met the prespecified non-inferiority criterion relative to SPAQ, with the upper bound of the 95% confidence interval remaining below the predefined margin. Although the observed hazard ratios suggested lower malaria incidence in the DHAPQ arm, the trial was primarily designed as a non-inferiority study. Nevertheless, the consistency of protective effects across transmission seasons and age groups supports the potential utility of DHAPQ in extended-age SMC strategies. Results from the intention-to-treat analysis, including additional comparisons relative to the control arm, were consistent with the primary per-protocol findings and are provided in Supplementary Figure S2.

Both extended-age regimens reduced asymptomatic parasitemia, but DHAPQ achieved the largest and most consistent reductions across years and age groups. Odds of infection in the DHAPQ arm were 20–70% lower than in the Control arm, highlighting its strong suppressive effect on parasitemia. These findings align with previous studies in Burkina Faso and Mali demonstrating substantial reductions in parasitemia following SMC.^12,26 Given the growing recognition that school-aged children constitute the primary reservoir sustaining transmission, the ability of DHAPQ to reduce asymptomatic parasitemia in this group has major implications for malaria control.

Our results are consistent with a study conducted in Uganda in young children, which showed that DHAPQ, when used for monthly prophylaxis, is well tolerated and significantly reduces the incidence of malaria.²⁷ The strong performance of DHAPQ across two seasons and two age groups underscores its potential as a more effective SMC regimen in high-transmission settings.

Although extension of SMC to older children has been explored in several Sahelian settings, including Senegal and Burkina Faso, implementation has remained limited, and evidence has largely been derived from non-randomized or single-setting studies. Our findings therefore, do not introduce the concept of age extension per se, but provide robust, comparative, multi-season evidence from a cluster-randomized design,²⁸ integrating epidemiological, entomological, and implementation outcomes within a single framework. This strengthens the evidence base for policy decisions in settings where malaria burden is shifting toward older children. Unlike the Uganda trial,²⁹ which showed non-inferiority of DHAPQ without evidence of superiority, our findings indicate greater protection in older children under high-transmission Sahelian conditions, though differences in study populations and epidemiological contexts limit direct comparability.

Entomological indicators revealed intense and sustained transmission across all study arms, with peak EIRs in September and October. Although DHAPQ villages exhibited higher EIRs (reflecting ecological rather than intervention differences), clinical malaria incidence was lowest in the DHAPQ arm. The highest EIR observed in the DHAPQ arm may have attenuated the protective effect estimates and, represent a residual confounding that cannot be completely excluded. However, the divergence between entomological risk and clinical outcomes highlights the strong protective efficacy of DHAPQ even under high transmission pressure.

Molecular surveillance conducted within a parallel study from the same trial did not identify evidence of increasing piperaquine resistance markers.²³ However, the detection of rare PfK13 mutations and emerging dhps variants highlights the importance of continued molecular surveillance, particularly if DHAPQ use expands programmatically.¹¹

These findings suggest that DHAPQ may represent a promising option for extended-age SMC strategies, with potential advantages in acceptability, coverage, and tolerability. However, in children under five years, DHAPQ was non-inferior but not consistently superior to SPAQ and therefore, does not provide sufficient evidence to support replacement of SPAQ in the current target population. Molecular data from a parallel study conducted within the same trial in Mali did not indicate an increase in markers associated with piperaquine resistance over the two-year period 25. Nevertheless, the detection of rare PfK13 mutations and emerging dhps variants, together with the relatively short duration of follow-up, underscores the need for continued molecular surveillance, particularly given concerns about artemisinin resistance in sub-Saharan Africa and the theoretical risk that widespread ACT-based chemoprevention could accelerate resistance selection. Cost considerations and operational feasibility must also be carefully weighed. Taken together, these findings support further evaluation of DHAPQ in expanded SMC strategies, while emphasizing that policy decisions should remain cautious and context-specific, considering efficacy, resistance dynamics, cost, and operational feasibility this study has some limitations. First, only the first dose of each SMC round was directly observed, meaning coverage estimates reflect day-1 administration rather than full three-day adherence.

Second, the open-label design may have introduced performance bias, particularly in caregiver reporting of adherence and adverse events. Third, malaria incidence was measured through passive case detection, which may underestimate true incidence and introduce variability in outcome ascertainment. To mitigate this, trained study clinicians were deployed at health facilities, and caregivers were encouraged to seek care for febrile illness. Fourth, a sensitivity analysis using a parasite density threshold (e.g., ≥5000 parasites/μL) was not conducted to define clinical malaria, consistent with previous SMC trials, but still our definition aligns with routine clinical malaria in Malian settings. Fifth, analytical limitations include reliance on cluster-level randomization with a small number of clusters, potential residual confounding, and limited power to detect interaction effects. To address this, analyses incorporated cluster-adjusted models, prespecified covariates, and adjustment for study year to account for the crossover design. And, finally, since this study used two well-known antimalaria drugs (already approved in SSA malaria endemic countries) SPAQ and DHAPQ, and for budgetary issues, side effect monitoring during the post-treatment period through house visits was not performed, but we rely on self-reported side effect and those observed only after the administration of the day 1 doses under DOT in both study arms. Another limitation of this study was the small number of clusters relative to their large sizes, which reduced our ability to fully account for clustering in a hierarchical model. This cluster-randomized crossover trial demonstrates that Dihydroartemisinine-piperaquine (DHAPQ) is non-inferior to SPAQ for SMC in children under five and provides substantially superior protection in children aged 5–9 years, a group increasingly recognized as a major reservoir of malaria transmission. DHAPQ was better tolerated, achieved higher and more stable coverage, and significantly reduced both asymptomatic parasitemia and clinical malaria incidence across two transmission seasons. These advantages, combined with growing concerns about SP and AQ resistance and declining acceptability of SPAQ, position DHAPQ as a strong candidate for next-generation SMC strategies in the Sahel. Large, multi-country implementation trials and cost-effectiveness analyses are now needed to inform policy decisions and support potential scale-up.

Research in context.

Evidence before this study

We searched PubMed through January 2026 using terms related to malaria, seasonal, malaria chemoprevention (SMC), and Mali, without language restrictions. Seventy-nine relevant articles were identified. Evidence indicates that although SPAQ remains effective, important challenges include persistent malaria among children aged 5–9 years, increasing SP/AQ resistance; and poor adherence due to adverse reactions. DHAPQ has shown promising efficacy and tolerability, but evidence is limited to single-season studies and lacks direct comparison with SPAQ in high-transmission Sahelian settings. Our study is the first multi-season, cluster-randomized trial in Mali comparing DHAPQ with SPAQ under routine SMC delivery, generating evidence to inform extended-age SMC and alternative chemoprevention regimens.

Added value of this study

This two-year cluster-randomized crossover non-inferiority trial involving 12,854 children, in Mali, provides the first multi-season comparison of extended-age SMC and an alternative drug regiment in a high-transmission Sahelian setting. Extending SMC to children aged 5–9 years substantially clinical malaria and asymptomatic parasitemia. DHAPQ was non-inferior to SPAQ in children under five and provided greater protection in school-aged children. By integrating epidemiological, entomological, implementation, and molecular data, the study generates comprehensive evidence to inform SMC policy and future resistance monitoring.

Implications of all the available evidence

Available evidence supports extending SMC to children aged 5–9 years in settings with persistent malaria burden, DHAPQ represents a promising alternative to SPAQ, but policy decisions should consider resistance, cost, operational feasibility, and long-term effectiveness, supported by continued surveillance and further evaluation.

Appendix A. Supplementary data

Supplementary data related to this article can be found at https://doi.org/10.1016/j.lanafr.2026.100122.

Data sharing statement

The protocol and data supporting the findings of this study will be made available upon reasonable request to the corresponding author.