Asymptomatic malaria: a hidden reservoir for the transmission and circulation of artemisinin partial-resistant Plasmodium falciparum in Tanzania

ABSTRACT

Asymptomatic malaria infections contribute substantially to silent transmission, but the prevalence of artemisinin resistance (ART-R) markers in these carriers remains poorly understood. A community-based cross-sectional study was conducted in Tanzania from December 2022 to July 2023, enrolling 3,489 participants from high-transmission regions of Geita and Kigoma and a low-transmission region of Arusha. Four villages per region were randomly selected, and venous blood samples were tested using rapid diagnostic tests, microscopy, and qPCR, revealing overall positivity rate of 24.4%, 15.8%, and 26.2%, respectively, which indicate a significant proportion of submicroscopic infections. Among the 802 isolates successfully sequenced for pfk13 and pfmdr1, 24 (3.0%) isolates from high-transmission areas carried validated pfk13 partial-resistance markers Y493H (0.2%), R561H (2.0%), and A675V (0.7%), while all low-transmission isolates were wild-type. All isolates retained the pfmdr1 N86 codon, and the NFD haplotype associated with reduced susceptibility to lumefantrine was detected in 48.1% and 48.4% of isolates in high- and low-transmission areas, respectively. Mutations were more frequent in children under five and in females. Artemether-lumefantrine (AL, 64.7%) was the most commonly used antimalarial in high-transmission areas, whereas sulfadoxine-pyrimethamine (SP, 75.9%) predominated in low-transmission areas. Higher AL use correlated with increased pfmdr1 mutation prevalence in high-transmission regions, while NFD detection in low-transmission areas may reflect gene flow from high-transmission settings. These findings demonstrate that asymptomatic carriers are a substantial hidden reservoir of ART-resistant parasites, emphasizing the importance of integrating molecular surveillance and demographic information on asymptomatic infections into malaria control programs to detect emerging resistance and guide targeted interventions in Tanzania.

Introduction

Malaria remains a leading cause of morbidity and mortality, particularly in sub-Saharan Africa (SSA), which accounts for the vast majority of global cases and deaths [1]. In 2022, an estimated 249 million malaria cases were reported globally, of which 94% (233 million) occurred in SSA [1]. In Tanzania, malaria distribution is highly heterogeneous, with annual incidence in children under five ranging from less than 1% in Arusha to 13% in Geita and Kigoma regions [2]. To address this burden, the Tanzanian government has committed to malaria elimination by 2030 through phased strategies to reduce transmission and mortality [3]. Preventive measures such as long-lasting insecticidal nets (LLINs), indoor residual spraying (IRS), and intermittent preventive treatment in pregnancy (IPTp) are widely implemented. However, these interventions do not specifically target asymptomatic infections. As a result, asymptomatic individuals may serve as hidden reservoirs, sustaining transmission and potentially facilitating the spread of artemisinin-resistant (ART-R) parasites [4]. Despite intensive control efforts, malaria cases in Tanzania declined only modestly from 4.9 million in 2015 to 4.5 million in 2021, indicating the continuing challenge [3].

“Asymptomatic malaria” refers to individuals infected with Plasmodium falciparum at any parasitaemia density, without fever or acute illness, and who have not received recent antimalarial treatment [5, 6]. These individuals act as silent reservoirs, contributing to ongoing transmission and potentially to the spread of drug-resistant parasites in the community [4]. Nevertheless, current National Malaria Control Programs (NMCP) largely overlook asymptomatic carriers, resulting in underestimation of their role in sustaining both transmission and antimalarial resistance [7].

Recent studies confirm pfk13 mutations associated with delayed parasite clearance in multiple SSA countries [4, 8]. Following widespread resistance to chloroquine (CQ) and sulfadoxine-pyrimethamine (SP), Tanzania adopted artemisinin-based combination therapy (ACT) as the first-line treatment for uncomplicated P. falciparum malaria. Among ACT regimens, Artemether-Lumefantrine (AL) is the preferred option, while alternative regimens such as Artesunate-Amodiaquine (ASAQ) and Dihydroartemisinin-Piperaquine (DHA-PPQ) are reserved for treatment failure or contraindications [9]. Injectable artesunate, an ART derivative, remains recommended for severe malaria. Recent therapeutic efficacy studies (TES) in Tanzania have reported high AL cure rates (> 95%), but declining efficacy has been noted in some neighbouring countries, raising concerns regarding the potential spread of resistant parasites [10, 11]. ACT efficacy depends on both the ART derivative and its partner drug. Therefore, the emergence of ART-R can reduce parasite clearance, compromise AL efficacy, and increase the risk of persistent infections and transmission, especially from asymptomatic carriers [12].

Genetic markers in the P. falciparum kelch 13 (pfk13) and P. falciparum multidrug resistance 1 (pfmdr1) genes are crucial for monitoring resistance to artemisinin and partner drugs [13, 14]. More than 260 non-synonymous pfk13 mutations have been reported, but only 13 residues are validated as ART-R markers, and nine are classified as candidate markers by the WHO [15]. Recent clinical studies in African countries have identified emerging pfk13 mutations in symptomatic cases. These include validated markers such as R561H and candidate markers like R622I, C469Y, and A675V [16-19]. In parallel, pfmdr1 polymorphisms, particularly at codon N86Y and Y184F, have been linked to reduced susceptibility to lumefantrine, the ACT partner drug, and may influence treatment outcomes [20]. Monitoring both pfk13 and pfmdr1 mutations is therefore critical to assess the potential spread of ART and partner drug resistance in endemic populations.

Asymptomatic P. falciparum infections, often submicroscopic, are increasingly reported in school-age children in some endemic areas, with prevalence reaching 20–40% [7]. Despite their prevalence, few studies have assessed whether these infections carry ART-R or partner drug resistance markers, highlighting a critical gap in surveillance strategies [21, 22]. Most existing data on pfk13 and pfmdr1 mutations in Tanzania come from symptomatic patients enrolled in therapeutic efficacy studies (TES), leaving asymptomatic reservoirs largely uncharacterized. As the country moves toward elimination, understanding whether these silent infections sustain or disseminate resistant parasites is essential to prevent resurgence and safeguard ACT efficacy. Given the scarcity of data on asymptomatic carriers in Tanzania, strengthening surveillance across both symptomatic and asymptomatic populations is critical [23]. This study investigates the prevalence of ART-R and partner drug resistance markers among asymptomatic carriers across diverse transmission settings in Tanzania. Specifically, it compares the prevalence and distribution of pfk13 and pfmdr1 polymorphisms between high and low-transmission areas and across different demographic groups. The findings are expected to provide critical evidence to guide malaria control strategies that address both symptomatic and asymptomatic infections, support Tanzania’s national elimination agenda by 2030, and inform broader regional and global efforts.

Materials and methods

Study design and setting

This was a community-based cross-sectional study conducted in Tanzania from December 2022 to July 2023. A multi-stage sampling method was employed to select the study sites. Three regions were purposively selected to represent both high (Kigoma and Geita) and low (Arusha) malaria transmission settings. This design allowed a comparative analysis of asymptomatic P. falciparum infections and prevalence of ART-R markers across diverse epidemiological settings. According to the 2022 Tanzania Demographic and Health Survey and Malaria Indicator Survey (DHS/MIS), malaria prevalence in under-fives was 13% in both Kigoma and Geita and less than 1% in Arusha [2]. Within each region, two districts were randomly selected, followed by a random selection of two villages from each district (Figure 1). In total, the study covered 12 villages across Tanzania, with eight villages from high transmission areas and four villages from low transmission areas (Figure 2). This selection ensured the survey reflects different transmission settings in Tanzania and provided a representative overview for assessing ART-R among asymptomatic carriers.

Figure 1.

Map of Tanzania. Map of Tanzania showing the study regions (Geita and Kigoma for high-transmission settings, and Arusha for low-transmission settings) and districts (Geita: Chato and Nyang’hwale; Kigoma: Kasulu and Kibondo; Arusha: Arusha DC and Meru). Colour patterns indicate the prevalence of malaria cases per 1,000 children under 5 years in 2022, as reported by the Tanzania Demographic and Health Survey and Malaria Indicator 2022–Key Indicators Report. The map was created using QGIS 3.34.1.

Figure 2.

Flow chart of study population. This flow chart shows the sample size (n) and proportions (%) at each study sites. High-transmission areas are highlighted in green, and low-transmission areas in yellow. qPCR + indicates the number and proportion of collected isolates that tested positive for P. falciparum. Among these, the number and proportion of samples successfully sequenced for both pfk13 and pfmdr1 (pfk13–pfmdr1 seq.) are shown.

Ethics statement

This study was approved by the National Institute for Medical Research (NIMR), a division of the Ministry of Health (MoH) in Tanzania (Approval number: NIMR/HQ/R.8a/Vol.IX/4114), and by the Institutional Review Board of Kangwon National University, Republic of Korea (Approval number: KWNUIRB-2022-06-008). All procedures were conducted in accordance with relevant national and international guidelines and regulations. Written informed consent was obtained from all participants or their legal guardians prior to enrolment.

Study population and eligibility criteria

Participants were individuals aged 1 year and older residing in the 12 selected villages. At the enrolment stage, individuals were excluded if they lived outside the selected areas, had used antimalarial drugs within the past 7 days, had life-threatening illnesses, or declined to provide informed consent. Children under 1 year were excluded due to technical challenges and ethical considerations associated with venipuncture in infants [24].

Screening for asymptomatic malaria included all household members present at the time of the visit, defined as afebrile (body temperature ≤ 37°C), with no malaria symptoms in the past 5 days, and no recent antimalarial use [25]. All collected blood samples were included in the diagnostic testing using rapid diagnostic tests (RDTs), light microscopy (LM), and quantitative PCR (qPCR), regardless of prior exclusion criteria. Sequencing of pfk13 and pfmdr1 genes was performed only on qPCR-positive samples. Demographic information and self-reported history of antimalarial use were collected using structured, pre-tested questionnaires administered by trained personnel.

Sample size calculation and representativeness

The sample size was calculated using a proportional formula, with a prevalence of 11.3% [26], a margin of error of 0.01, and a standard normal deviation of 1.96 corresponding to a 95% confidence interval. After accounting for a 10% non-response rate, the total sample size for the study was 3,360 participants, distributed across the 12 villages (resulting in an estimated sample size of 280 participants per village). Although the calculation was based on overall malaria prevalence, this sample size provides sufficient statistical power to detect low-prevalence ART-R markers (<1%) among malaria-positive individuals, supporting the survey’s ability to identify rare variants. This is consistent with previous regional surveillance in Tanzania, where PfK13-R561H prevalence ranged from 1.4% in low-transmission districts (Ngara) to 22.8% in high-transmission districts (Karagwe) [23]. Selecting regions across high- and low-endemic settings ensures the study captures the epidemiological diversity of malaria and potential ART-R prevalence in Tanzania.

Data collection and field procedures

The research team was guided by village chairpersons and community health workers. It included a trained medical doctor, laboratory technicians, consenters, and interviewers. They visited households to conduct the survey and collect samples. Houses were randomly selected from a list provided by the local administration offices. In each household, all members meeting inclusion criteria were invited to participate. Venous blood samples were collected for molecular testing, alongside demographic and clinical data. Participants were informed about the survey one day prior to the visit, and consent/assent procedures were strictly adhered to for minors.

Blood sample collection

For each participant who met the inclusion criteria, 3 mL of venous blood was collected into heparinized vacutainer tubes (BD Vacutainer® Plastic Heparin Tubes) using a sterile vacutainer syringe with a holder (BD Vacutainer® one-use holder). Immediately after collection, malaria was tested using a rapid diagnostic test (RDT) (Bioline™ Malaria Ag P.f/Pan: Abbott, Chicago, IL, USA) following the manufacturer’s guidelines. Participants who tested positive were treated with artemether-lumefantrine (AL), the recommended ACT according to Tanzania's malaria treatment policy.

A portion of the blood was transported on ice (2-8°C) to a nearby health facility, where blood smears were prepared and examined by two skilled microscopists. Both microscopists were certified by the Health Laboratory Practitioners Council (HLPC) of Tanzania and employed by the government as laboratory technicians in district hospitals where the study was conducted. When the results from the two primary microscopists were discordant, a third microscopist independently reviewed the slides, and their results were considered definitive.

The remaining whole blood was first spotted onto Whatman 903 Protein Saver Card (Cytiva, Seoul, Republic of Korea) to create five dried blood spots (DBS) per participant, each spot consisting of 50 µL of blood. DBS were allowed to dry overnight, packed separately into plastic bags with silica gel and transported to Kangwon National University (KNU) for genomic DNA (gDNA) extraction, molecular diagnosis by qPCR, and Sanger sequencing for ART and partner drug resistance marker genes. Another portion of the collected blood was then centrifuged to obtain serum for subsequent analyses, including antigenicity screening of target antigens in separate studies aimed at P. falciparum vaccine candidate validation [27].

DNA extraction and molecular diagnosis of P. falciparum infections

At Kangwon National University, a single spot (50 µL whole blood) from each DBS sample was cut into small pieces and placed into a 1.7 mL microcentrifuge tube. Parasite gDNA was then extracted using QIAamp DNA Mini Kit (Qiagen, Hilden, Germany), following the manufacturer’s instructions, and eluted in 50 µL of elution buffer. Quantitative PCR for the diagnosis of P. falciparum was performed by targeting the 18S ribosomal RNA gene, as previously described [28]. P. falciparum-specific primers (Forward: ATTGCTTTTGAGAGGTTTTGTTACTTT, Reverse: GCTGTAGTATTCAAACACAATGAACTCAA) and probe (FAM-CATAACAGACGGGTAGTCAT) were used for molecular diagnosis. The qPCR assays were optimized for study-specific conditions, with minor modifications to cycling parameters to ensure consistent and reliable amplification. Briefly, qPCR reactions were carried out in a total volume of 10 µL, comprising 5 µL of 2X Prime Time Gene Expression Master Mix with ROX reference dye (Integrated DNA technologies, Coralville, IA, USA), 1 µL of gDNA, 0.5 µM of each primer, and 0.25 µM of TaqMan Probe. Reactions were performed on an AriaMx Real-Time PCR system (Agilent, Santa Clara, CA, USA) using two-step cycling conditions: hot-start polymerase activation at 95℃ for 3 min, followed by 45 cycles of denaturation at 95℃ for 15 s and annealing/extension at 58℃ for 1 min.

A standard curve for qPCR diagnosis was prepared as previously described [29]. Briefly, the standard curve was generated using in vitro-cultured 3D7 strain parasites from KNU malaria laboratory. The gDNA of this positive control was extracted from parasites at 1% parasitaemia in the ring stage using QIAamp DNA Mini Kit (Qiagen). The standard curve was prepared using seven points of tenfold serial dilutions of gDNA that corresponded to 50,000 parasites/µL to 0.05 parasites/µL, with each dilution run in duplicate. In each reaction, a duplicate of distilled deionized water (DDW) was used as a negative control.

Amplification and sequencing of pfk13 and pfmdr1 from clinical isolates

Amplification of the pfk13 and pfmdr1 genes was performed on samples that were mono-infections of P. falciparum. Samples confirmed as P. falciparum infections by species-specific qPCR were selected for sequencing. Mixed-species infections were included, as they do not interfere with P. falciparum-specific PCR and sequencing, allowing for maximal recovery of sequence data. A target gene-specific nested PCR was conducted for each target of interest using established protocols and specific primer pairs (Table S1) on a ProFlex PCR System (Life Technologies, Singapore). Reactions were carried out in 20 µL volume using AccuPower® PCR PreMix (Bioneer, Daejeon, Republic of Korea). 1 µL of extracted P. falciparum gDNA was used as template for the first PCR and 1 µL of the first PCR product was used for the nested PCR reaction. All PCR conditions were as follows: pre-denaturation at 95 °C for 10 min, followed by 35 cycles of denaturing at 95°C for 30 s, annealing at 58°C for 30 s, and extension at 72°C for 30 s per 500 bp of target length, plus a final extension at 72°C for 10 min. Nested PCR products were purified using a DNA Purification Kit (Alphagen Biotech, Pingtung, Taiwan). Each PCR batch included gDNA from the P. falciparum 3D7 strain as a positive control and DDW (no-template control) as a negative control. The purified amplicons were confirmed using 1.2% agarose gel electrophoresis and subsequently subjected to Sanger sequencing using an ABI3730xl DNA Analyzer (Nbit, Chuncheon, Republic of Korea). For the pfk13 gene, the sequenced region covered the amino acid positions from codons 465–710. The pfmdr1 gene was sequenced in two regions: region 1 included codons 86 and 184, while region 2 covered the 1246 codon. Sequencing was conducted using specific primers (Table S1). Reference sequences for the P. falciparum 3D7 strain were retrieved from the PlasmoDB database (http://plasmodb.org/plasmo/) with the gene IDs: PF3D7_1343700 (pfk13) and PF3D7_0523000 (pfmdr1). The resulting sequence data have been deposited in the GenBank database under the accession numbers PX450997–PX451830 for pfk13 and PX449291–PX450087 for pfmdr1.

Data management, identification of SNPs and statistical analysis

To ensure proper data management and sampling, study sites were coded, and each participant was assigned a unique code corresponding to their site. During sample collection, these codes were matched to participants’ names using specialized forms. Only the codes were marked on blood samples and used throughout all processing steps. Participants’ codes, demographic details, and associated information were recorded and double-checked daily in an MS Excel file (Microsoft® 365).

Sanger sequencing data were analysed and visualized using SnapGene® 2.3.2 software (GSL Biotech, Chicago, IL). Multiple sequence alignments were performed to identify single nucleotide polymorphisms (SNPs) by referencing the P. falciparum 3D7. Chromatograms showing ambiguity or overlapping peaks were re-amplified and re-sequenced; unresolved cases were excluded from analysis to avoid potential bias from multiplicity of infections (MOI).

SNPs and haplotypes were calculated as the percentage of successfully sequenced samples. Differences in haplotype frequencies across age group and between genders were evaluated separately for each gene (pfk13 and pfmdr1) using Chi-square test or Fisher’s exact test, as appropriate. P-values were adjusted for multiple comparisons using Bonferroni correction. All statistical analyses were performed using GraphPad Prism 8.0.2, with the significance defined at P < 0.05 (95% confidence interval). No significant differences were observed after Bonferroni correction.

Results

Baseline characteristics and prevalence of asymptomatic malaria

A total of 3,489 samples were collected from asymptomatic individuals in the 12 selected villages (Figure 2). Asymptomatic malaria infections and demographic characteristics were confirmed in the field using RDT and LM, the performance of which has been described previously [7]. Among all participants, 2,365 (67.8%) were from high-transmission regions (Kigoma and Geita) and 1,124 (32.2%) were from low-transmission region (Arusha). Of these, 1,527 (43.8%) were males and 1,962 (56.2%) were females (Table 1). Of the 3,489 participants, 499 (14.3%) were younger than 5 years, 1,257 (36.0%) were aged 5–14 years, and 1,733 (49.7%) were 15 years or older. The median age was 14.0 years (interquartile range [IQR]: 7–36) (Table 1).

Table 1.

Baseline information of participants recruited in the study showing their number, proportions, study sites and transmission settings.

Variable	HIGH–TRANSMISSION STRATA								LOW–TRANSMISSION STRATA
	GEITA				KIGOMA				ARUSHA
	Chato		Nyang'hwale		Kasulu		Kibondo		Arusha DC		Meru
	Ihanga (n = 339)	Rwantaba (n = 278)	Kayenze (n = 327)	Nyangalamila (n = 276)	Mugombe (n = 290)	Nyamnyusi (n = 259)	Bunyambo (n = 303)	Kumuhasha (n = 293)	Bwawani (n = 284)	Themi ya simba (n = 279)	Maji ya chai (n = 279)	Ngurudoto (n = 282)
Sex, n (%)
Male	154 (45.4)	98 (35.2)	154 (47.1)	177 (64.1)	123 (42.4)	102 (39.4)	101 (33.3)	99 (33.8)	126 (44.4)	153 (54.8)	125 (44.8)	115 (40.8)
Female	185 (54.6)	180 (64.8)	173 (52.9)	99 (35.9)	167 (57.6)	157 (60.6)	202 (66.7)	194 (66.2)	158 (55.6)	126 (45.2)	154 (55.2)	167 (59.2)
Age in years
Age group, n (%)
< 5	50 (14.8)	65 (23.4)	53 (16.2)	37 (13.4)	26 (9.0)	33 (12.7)	76 (25.1)	49 (16.7)	36 (12.7)	34 (12.2)	20 (7.2)	20 (7.1)
5 - 14	144 (42.5)	122 (43.9)	81 (24.8)	98 (35.5)	172 (59.3)	99 (38.2)	136 (44.9)	135 (46.1)	88 (31.0)	85 (30.5)	38 (13.6)	59 (20.9)
≥ 15	145 (42.8)	91 (32.7)	193 (59.0)	141 (51.1)	92 (31.7)	127 (49.0)	91 (30.0)	109 (37.2)	160 (56.3)	160 (57.3)	221 (79.2)	203 (72.0)

Malaria prevalence in high-transmission regions was 38.0%, 20.7%, and 34.8% in Geita (n = 1,220) and 33.1%, 25.1%, and 38.7% in Kigoma (n = 1,145) by RDT, LM, and qPCR, respectively (Figure 3A). In the low-transmission region of Arusha (n = 1,124), prevalence was markedly lower at 2.8%, 1.1%, and 4.2% across the same diagnostic methods (Figure 3A). Malaria prevalence was significantly lower in low-transmission settings across all three diagnostic methods (p < 0.001), showing consistent trends among assays (Figure 3A). Overall, among the 3,489 study participants, positivity rates were 24.4% (n = 853), 15.8% (n = 552), and 26.2% (n = 915) by RDT, LM and qPCR, respectively (Figure 3A).

Figure 3.

Baseline information of the study population. (A) Prevalence of asymptomatic malaria in selected regions based on diagnostic assays, including rapid diagnostic test (RDT), Light microscopy (LM), and qPCR. The black bars indicate the proportion of positive (+) individuals (%), and the yellow bars indicate the proportion of negative (–) individuals (%). (B) Distribution of P. falciparum–positive isolates detected by LM, RDT and qPCR, and selected for genotyping and sequencing (qPCR positive) of the pfk13 and pfmdr1 genes. (C) Distribution of WHO-validated pfk13 mutations associated with partial artemisinin resistance (A675 V, R561H, and Y493H) among asymptomatic populations in the studied villages. (D) Distribution of pfmdr1 SNPs linked to multidrug resistance markers known to modulate the susceptibility of P. falciparum parasites to multiple drugs, including lumefantrine.

The 915 (26.2%) participants who tested positive by qPCR were selected for pfk13 and pfmdr1 polymorphism analysis (Figure 3B). SNP analysis was successfully performed on 834 isolates for pfk13 and 804 isolates for pfmdr1. A total of 802 isolates (771 from high-transmission and 31 isolates from low-transmission) were successfully sequenced for both genes, yielding an overall sequencing success rate of 87.7% across transmission strata (Figure 2).

Molecular marker polymorphisms in the pfk13 gene

For the pfk13 gene, 834 out of 915 (91.1%) isolates were successfully sequenced. Among these, 802 isolates were from high malaria transmission villages, while 32 isolates originated from low transmission villages (Table S2). The majority of sequenced isolates, 794 (95.2%), harboured pfk13 wild-type parasites, while 23 (2.8%) carried WHO-validated pfk13 markers associated with partial resistance to artemisinin. These markers were exclusively found in isolates from high transmission villages and were present at codons Y493H (n = 2, 0.2%), R561H (n = 15, 1.8%), and A675V (n = 6, 0.7%) (Table S2 and Figure 3C). Additionally, 17 isolates (2.0%) exhibited non-synonymous mutations that are not currently validated as ART-R markers (Table S2). These mutations occurred at codons V487I, I501V, S522I, N554T, N554S, W565R, A569P, A569V, A578S, D648V, and F662S. Furthermore, 10 (1.2%) isolates displayed synonymous mutations at codons G497G, T508T, L524L, R528R, R597R, E567E, I601I, A626A, and N629N.

Molecular marker polymorphisms in the pfmdr1 gene

The pfmdr1 gene was sequenced to determine the prevalence of the N86 wild-type codon and the Y184F and D1246Y mutant codons, which have been previously reported to modulate parasite susceptibility to partner drugs of ART, including lumefantrine, the partner drug of ACT in Tanzania. A total of 804 (87.9%) samples were successfully sequenced for both the first and second pfmdr1 fragments (Table S3). The first fragment encompassed codons 86 and 184, while the second fragment included codon 1246. No mutations were observed at codon N86, which remained wild-type in all sequenced isolates. In high-transmission settings, the prevalence of the Y184F variant was significantly higher in Geita (226/378; 59.8%) compared to Kigoma (149/393; 37.9%) (p < 0.001). Despite the smaller number of samples sequenced from Arusha, the mutant Y184F codon was prevalent in 15 of 33 (45.5%) isolates. The D1246Y mutation was identified in only one isolate from Bwawani village in Arusha (1/12; 8.3%) (Figure 3D). This corresponds to 3.0% of the 33 isolates in this low-transmission settings. In high-transmission settings (Geita and Kigoma), 8 of 771 (1.0%) isolates sequenced from five of the eight villages harboured the D1246Y mutant codon. The distribution of the 184F and 1246Y mutant codons varied significantly across the studied villages (Figure 3D). Overall, four haplotypes were classified as wild-type NYD (410 isolates, 51.0%), single mutant NFD (385 isolates, 47.9%), single mutant NYY (4 isolates, 0.5%), and double mutant NFY (5 isolates, 0.6%) (Table S3).

Demographic distribution of different pfk13 and pfmdr1 SNPs and haplotypes

A total of 802 fully sequenced isolates from 12 villages were analysed for markers in pfk13 and pfmdr1 (Table 2). Nine distinct pfk13-pfmdr1 haplotypes were identified. The most prevalent haplotype was wild-type pfk13 combined with single-mutant pfmdr1-184F (YRA-NFD), present in 374 isolates (46.6%). Eleven isolates from high-transmission villages harboured pfk13 single mutations associated with ART-R, while another 11 isolates exhibited co-occurrence of pfk13 and pfmdr1 mutations, which are known to reduce susceptibility to AL (Table 2).

Table 2.

Co-prevalence of the different pfk13 (Y493H, R561H and A675V) and pfmdr1 (N86Y, Y184F and D1246Y) SNPs and haplotypes.

Transmission profile	District	Region	Village		PfK13-PfMDR1 SNPs and haplotypes, n (%)
				Wild-type YRA-NYD	Sensitive* YRA-NFD	Sensitive YRA-NYY	Sensitive YRA-NFY	ART-R** YHA-NYD	ART-R YRV-NYD	AL-R*** HRA-NFD	AL-R YHA-NFD	AL-R YRV-NFD
High transmission	Geita	Chato	Ihanga (n = 111)	34 (30.6)	74 (66.7)	–	–	–	–	–	2 (1.8)	1 (0.9)
			Rwantaba (n = 112)	56 (50.0)	50 (44.6)	2 (1.8)	–	–	3 (2.7)	–	–	1 (0.9)
		Nyang’hwale	Kayenze (n = 79)	16 (20.3)	57 (72.1)	–	3 (3.8)	–	–	2 (2.5)	1 (1.3)	–
			Nyangalamila (n = 76)	41 (53.9)	35 (46.1)	–	–	–	–	–	–	–
	Kigoma	Kasulu	Mugombe (n = 109)	64 (58.7)	43 (39.5)	–	–	–	1 (0.9)	–	1 (0.9)	–
			Nyamnyusi (n = 96)	65 (67.7)	28 (29.2)	2 (2.1)	1 (1.0)	–	–	–	–	–
		Kibondo	Bunyambo (n = 97)	61 (62.9)	34 (35.1)	1 (1.0)	1 (1.0)	–	–	–	–	–
			Kumuhasha (n = 91)	41 (45.1)	38 (41.8)	–	–	9 (9.9)	–	–	3 (3.3)	–
			Sub total (n = 771)	378 (49.0)	359 (46.6)	5 (0.6)	5 (0.6)	9 (1.2)	4 (0.5)	2 (0.3)	7 (0.9)	2 (0.3)
Low transmission	Arusha	Arusha DC	Bwawani (n = 14)	6 (42.9)	7 (50.0)	1 (7.1)	–	–	–	–	–	–
			Themi ya simba
(n = 9)	5 (55.6)	4 (44.4)	-	–	–	–	–	–	–
		Meru	Maji ya chai (n = 5)	2 (40.0)	3 (60.0)	–	–	–	–	–	–	–
			Ngurudoto (n = 5)	4 (80.0)	1 (20.0)	–	–	–	–	–	–	–
			Sub total (n = 31)	15 (48.4)	15 (48.4)	1 (3.2)
			TOTAL (n = 802)	393 (49.0)	374 (46.6)	6 (0.7)	5 (0.6)	9 (1.1)	4 (0.5)	2 (0.2)	7 (0.9)	2 (0.2)

*Indicate isolates with mutant SNPs in both pfk13 and pfmdr1 genes.

Bold letters indicate mutant codons.

The prevalence of mutant SNPs in pfk13 (Y493H, R561H and A675V) and pfmdr1 (N86Y, Y184F, and D1246Y) tended to be higher among children under five years of age and gradually declined with increasing age, although these differences did not reach statistical significance, they suggested a trend toward higher mutation frequency in younger individuals (Figure 4A).

Figure 4.

Co-prevalence of pfk13 and pfmdr1 SNPs and haplotypes. Distribution of mutant variants associated with antimalarial drug susceptibility in the pfk13 and pfmdr1 genes across (A) age groups and (B) gender groups of participants. The SNPs of pfk13 and pfmdr1 were unevenly distributed among age groups, with a higher prevalence observed in children under five years of age and in females. The proportions of wild-type (3D7 strain) variants are also indicated. (C) Co-prevalence and distribution of SNPs and haplotypes in the pfk13 and pfmdr1 genes across the studied regions. The co-prevalence of the pfk13 wild-type haplotype and the pfmdr1-NFD haplotype was significantly higher and more widespread in all villages in both transmission settings.

Mutant SNPs in pfk13 and pfmdr1 were more frequent in females (3.2% and 57.5%, respectively) than in males (0.8% and 34.5%, respectively) (pfk13, p = 0.02; pfmdr1, p < 0.0001), suggesting sex-specific differences in the distribution of these mutations (Figure 4B).

Mutations in pfk13 and pfmdr1 displayed marked spatial heterogeneity across the study villages. The proportion of wild-type isolates ranged from 20.3% in Kayenze (high-transmission village) to 80.0% in Ngurudoto (low-transmission village), reflecting variation in mutation distribution across transmission setting (Figure 4C).

Antimalarials use and its association with the prevalence of drug resistance markers

Participants provided self-reported histories of antimalarial use through structured questionnaires. The analysis highlighted differences in drug use behaviour, preferences, and sources. Among 2,365 participants from high-transmission regions, 153 (6.5%) reported antimalarial use beyond the seven-day exclusion period, compared to 29 of 1,124 (2.6%) participants from low-transmission regions (Table 3). Artemether-lumefantrine, the first-line ACT in Tanzania, was the most frequently used antimalarial. Drug use patterns differed significantly between transmission strata (Fisher’s exact test, p < 0.0001), with 99 of 153 (64.7%) participants in high-transmission regions and 7 of 29 (24.1%) participants in low-transmission regions reporting AL use (Table 3). In contrast, 13 (8.5%) participants from high-transmission regions reported using traditional medicines, whereas no traditional medicine use was reported in low-transmission regions (Table 3). The sources of drugs varied, with pharmacies (75 participants, 41.2%) and health facilities (53 participants, 29.1%) being the most common, followed by leftover medications (34 participants, 18.7%) and traditional medicines (11 participants, 6.0%) (Table 3).

Table 3.

Antimalarial uses, sources and association to mutant markers within selected villages.

Transmission Profile	District	Region	Village (n*)	Antimalarials (n, %)							Sources of drugs (n, %)						Mutant SNPs (n, %)
				AL	SP	CQ	AMQ	QN	TM	OTH	HF	TH	PH	TM	LO	OTH	Pf + ve	pfk13	pfmdr1
High transmission	Geita	Chato	Ihanga (n = 15)	11 (73.3)	1 (6.7)	–	–	–	–	3 (20.0)	3 (20.0)	–	6 (40.0)	3 (20.0)	3 (20.0)	–	4 (26.7)	1 (25.0)	2 (50.0)
			Rwantaba (n = 15)	10 (66.7)	5 (33.3)	–	–	–	–	–	4 (26.7)	–	9 (60.0)	0 (0.0)	2 (13.3)	–	8 (53.3)	1 (12.5)	4 (50.0)
		Nyang’hwale	Kayenze (n = 22)	16 (72.7)	3 (13.6)	–	–	–	3 (13.6)	–	3 (13.6)	–	12 (54.5)	2 (9.1)	5 (22.7)	–	19 (86.4)	2 (10.5)	15 (78.9)
			Nyangalamila (n = 18)	9 (50.0)	5 (27.8)	–	–	–	3 (16.7)	1 (5.6)	6 (33.3)	–	8 (44.4)	–	4 (22.2)	–	4 (22.2)	–	3 (75.0)
	Kigoma	Kasulu	Mugombe (n = 19)	11 (57.9)	5 (26.3)	–	–	–	3 (15.8)	–	4 (21.1)	–	10 (52.6)	1 (5.3)	4 (21.1)	–	8 (42.1)	1 (12.5)	6 (75.0)
			Nyamnyusi (n = 18)	10 (55.6)	6 (33.3)	–	–	–	2 (11.1)	–	6 (33.3)	–	6 (33.3)	2 (11.1)	4 (22.2)	–	4 (22.2)	–	2 (50.0)
		Kibondo	Bunyambo (n = 22)	18 (81.8)	4 (18.2)	–	–	–	–	–	7 (31.8)	–	7 (31.8)	–	5 (22.7)	3 (13.6)	4 (18.2)	1 (25.5)	2 (50.0)
			Kumuhasha (n = 24)	14 (58.3)	5 (20.8)	–	–	–	1 (4.2)	4 (16.7)	8 (33.3)	1 (4.2)	5 (20.8)	3 (12.5)	3 (12.5)	4 (16.7)	12 (50.0)	–	9 (75.0)
			Sub-total (n = 153)	99 (64.7)	34 (22.2)	–	–	–	12 (7.8)	8 (5.2)	41 (26.8)	1 (0.7)	63 (41.2)	11 (7.2)	30 (19.6)	7 (4.6)	63 (41.2)	6 (9.5)	43 (68.3)
Low transmission	Arusha	Arusha DC	Bwawani (n = 12)	3 (25.0)	9 (75.0)	–	–	–	–	–	7 (58.3)	–	2 (16.7)	–	3 (25.0)	–	2 (16.7)	–	1 (50.0)
			Themi ya simba (n = 5)	1 (20.0)	4 (80.0)	–	–	–	–	–	3 (60.0)	–	2 (40.0)	–	–	–	–	–	–
		Meru	Maji ya chai (n = 8)	2 (25.0)	6 (75.0)	–	–	–	–	–	–	–	6 (75.0)	–	1 (12.5)	1 (12.5)	–	–	–
			Ngurudoto (n = 4)	1 (25.0)	3 (75.0)	–	–	–	–	–	2 (50.0)	–	2 (50.0)	–	–	–	1 (25.0)	–	–
			Sub-total (n = 29)	7 (24.1)	22 (75.0)	–	–	–	–	–	12 (41.4)	–	12 (41.4)	–	4 (13.8)	1 (3.4)	3 (10.3)	–	1 (33.3)
			TOTAL (n = 182)	106 (58.2)	56 (30.8)	–	–	–	12 (6.6)	8 (4.4)	53 (29.1)	1 (0.5)	75 (41.2)	11 (6.0)	34 (18.7)	8 (4.4)	66 (36.3)	6 (9.1)	44 (66.7)

*n, of drug use.

Antimalarials (AL = Artemether-lumefantrine, SP = Sulfadoxine-pyrimethamine, CQ = Chloroquine, AMQ = Amodiaquine, QN = Quinine, TM = Traditional Medicines, OTH = Others).

Sources of drugs (HF = Health Facility, TH = Traditional Healers, PH = Pharmacies, TM = Traditional Medicines, LO = Leftovers, OTH = Others).

Among the 182 participants with a history of drug use, 63 of 153 (41.2%) cases from high-transmission regions and 3 of 29 (10.3%) cases from low-transmission regions tested positive for P. falciparum (Table 3). Sequencing of the isolates for pfk13 and pfmdr1 mutations revealed that 6 (9.1%) and 43 (68.3%) of 63 isolates from high-transmission villages carried mutant SNPs in pfk13 and pfmdr1, respectively. In low-transmission villages, only one participant carried a pfmdr1 mutant SNP, and all isolates harboured the wild-type codon at the pfk13 locus (Table 3). These results indicate a higher prevalence of pfk13 and pfmdr1 mutations in high-transmission villages, consistent with drug pressure, particularly from AL, which may select for P. falciparum parasites carrying drug-tolerance markers in pfmdr1 or partial ART-R markers in pfk13.

Discussion

With the global goal of malaria elimination by 2030, addressing asymptomatic reservoirs of artemisinin resistance (ART-R) has emerged as a critical priority in malaria-endemic regions [9]. This concern is particularly relevant in sub-Saharan Africa, where asymptomatic carriers constitute a major fraction of the parasite reservoir and transmission intensity varies considerably across regions [7, 15, 30]. Despite their potential impact, systematic evaluations of ART-R in asymptomatic infections remain limited, and their epidemiological implications remain poorly understood [12]. Asymptomatic and submicroscopic infections can harbour transmissible gametocytes, thereby sustain local transmission and underscoring their public health importance [31, 32]. Thus, addressing this knowledge gap is essential for designing effective surveillance strategies and targeted interventions aimed at both controlling transmission and preventing the emergence and spread of drug-resistant parasites.

Molecular surveillance in this study provides evidence that asymptomatic infections in Tanzanian villages carry validated ART-R markers and partner drug resistance polymorphisms, confirming the presence of silent reservoirs that could compromise ACT efficacy if unmonitored. Several pfk13 propeller domain mutations associated with delayed parasite clearance have been reported in symptomatic cases across African countries, including Rwanda, Uganda, and Ghana [19, 33-36]. In Tanzania, previous studies have shown heterogeneous distributions of R561H, A675V, and Y493H in both symptomatic and asymptomatic infections [18]. Consistent with these reports, R561H was the most prevalent mutation in high-transmission villages (0.9–11.0%), whereas A675V and Y493H were detected only in limited locations. The detection of Y493H in Tanzania for the first time, even at low frequency, suggests the emergence of ART-R parasites and underscores the need for ongoing surveillance in asymptomatic carriers. These findings are in line with nationwide molecular surveillance data, which also reported variation in R561H frequencies across transmission strata, ranging from 1.4% in Ngara and 22.8% in Karagwe [18, 23], thereby confirming the persistence of this allele in high-transmission settings. This geographic heterogeneity suggests that local transmission intensity, ACT usage, and host immunity contribute to the selection and maintenance of specific pfk13 alleles [37]. Compared with symptomatic malaria cases, asymptomatic infections generally exhibited lower frequencies of pfk13 mutations, as reported in several African and Southeast Asian studies [37, 38]. This pattern may reflect differences in antimalarial drug exposure between infection types. Taken together, these observations suggest that pfk13 mutations in asymptomatic carriers may represent an early warning signal for emerging resistance rather than isolated genetic drift, particularly when observed alongside partner drug resistance markers [19, 39, 40].

Patterns of antimalarial use further contextualize these molecular findings. Drug exposure was higher in high-transmission strata, with AL being the predominant antimalarial (64.7%) compared with low-transmission strata (24.1%), corresponding to a higher prevalence of pfmdr1 mutations in high-transmission strata (68.3%) compared with low-transmission strata (33.3%) among P. falciparum-positive individuals who had used antimalarials. Overall, molecular surveillance of pfmdr1 revealed that all isolates sequenced at codon 86 carried the wild-type allele, and the NFD haplotype (N86, 184F, 1034D) was predominant across study villages. This predominance of the NFD haplotype under AL selection pressure has been evident in several other African countries, suggesting that widespread AL use exerts similar selective pressure on pfmdr1 across diverse endemic settings and may promote regional convergence in drug-resistance dynamics [11, 41-43]. The persistence of NFD haplotypes suggests that ongoing ACT pressure may favour multidrug-tolerant parasites, similar to patterns seen with pfk13 [23]. These patterns are consistent with recent TES data from Tanzania, where AL efficacy remains generally high but varies considerably across regions (78–95%) [10, 11, 44]. Similar heterogeneity has been documented across Africa (I² = 95.9%), ranging from ≥95% in Ivory Coast, Mozambique, and Nigeria to <80% in Ethiopia, Burkina Faso, and Uganda, with intermediate values (80–90%) reported in Tanzania, Ghana, and Kenya [11]. These regional differences likely reflect variation in transmission intensity, treatment pressure, and distribution of pfmdr1 haplotypes associated with lumefantrine tolerance. While current TES results indicate sustained AL efficacy, the observed persistence of pfk13-R561H and pfmdr1-NFD in asymptomatic infections could signify early-stage multidrug tolerance, warranting proactive containment measures before clinical resistance becomes widespread. These findings support the implementation of targeted interventions such as focal mass drug administration (MDA) in high-transmission areas, together with continued therapeutic and molecular surveillance to mitigate further resistance spread.

Importantly, host and behavioural factors were also found to influence ART-R dynamics. Children under five exhibited higher proportions of mutant haplotypes in both pfk13 and pfmdr1, although these differences were not statistically significant. This trend is consistent with their lower acquired immunity and similar findings from other Tanzanian and African paediatric cohorts [45-47]. Additionally, this pattern may reflect higher ACT exposure in younger children and suggests that age-targeted interventions could be considered to reduce resistance selection pressure. Females showed a higher prevalence of resistance-associated polymorphisms compared with males. This sex-stratified pattern may reflect biological differences, such as sex-specific immune responses to P. falciparum, as well as epidemiological and behavioural factors, including differential access to antimalarials, adherence patterns, exposure risks, and socio-cultural practices [7, 48, 49]. Variability in drug sources, including pharmacies, leftover medications, and traditional remedies, may expose parasites to sub-therapeutic drug levels, further supporting the emergence of resistant strains [50, 51].

Several limitations of this study should be noted. Its cross-sectional design precludes inference of causality or temporal trends. In addition, as sampling was conducted between December and July, potential seasonal bias in parasite prevalence and allele frequency cannot be excluded. Reliance on self-reported drug use and absence of phenotypic validation may introduce measurement bias. Despite these limitations, the study provides comprehensive molecular insights into asymptomatic ART-R reservoirs, informing targeted interventions such as focal MDA or intensified surveillance in high-transmission villages. Such strategies, guided by molecular and TES data, can optimize resource allocation and curb the spread of resistant parasites.

Taken together, asymptomatic P. falciparum infections in Tanzania carry both ART-R and partner drug resistance markers, representing silent reservoirs that can sustain the spread of resistant parasites. These findings emphasize the strategic importance of surveillance targeting asymptomatic carriers to detect and contain emerging resistance. Integrating molecular data with TES and community-level epidemiological information will be essential for guiding evidence-based malaria control strategies, preserving ACT efficacy, and supporting progress toward the global malaria elimination goal by 2030.

Conclusions

This study highlights the complex dynamics of antimalarial resistance, emphasizing the critical role of asymptomatic carriers, community-level disparities, antimalarial drug use, transmission intensity, and demographic factors such as age and gender in shaping resistance patterns. The detection of partial artemisinin resistance mutations and pfmdr1 variants linked to reduced antimalarial susceptibility underscores the need for targeted surveillance among asymptomatic individuals to curb the silent spread of resistant parasites. Regional heterogeneity in pfk13 and pfmdr1 mutations suggests that selective pressure from widespread antimalarial use is a key driver of resistance evolution. Moreover, the greater vulnerability of young children and the observed gender differences highlight the importance of inclusive, context-specific control strategies. Achieving the global malaria elimination goal by 2030 will require the integration of robust molecular surveillance, therapeutic efficacy studies, and community-focused interventions to ensure equitable and effective management of drug resistance across diverse populations.

Funding Statement

This study was supported by the Korea Health Technology R&D Project through the Korea Health Industry Development Institute (KHIDI), funded by the Ministry of Health & Welfare [HI22C0820 & RS-2025-02309009], the National Research Foundation of Korea (NRF) funded by the Ministry of Education [RS-2023-00240627], and Ministry of Science and ICT [RS-2025-16069701] (J-H. H.).

Disclosure statement

No potential conflict of interest was reported by the authors.

Supplemental Material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/22221751.2025.2602320