Malaria is a mosquito-transmitted parasitic disease that occurs primarily in impoverished tropical and subtropical areas of the world. In the Greater Mekong Subregion (GMS), which consists of Cambodia, China’s Yunnan and Guangxi provinces, the Lao People’s Democratic Republic (Laos), Myanmar, Thailand, and Vietnam, malaria has been one of the most severe public health issues, hampering socioeconomic development.1–3 Recent decades have welcomed bourgeoning economic growth and significant improvement in public health in GMS countries. Driven by increasing political commitment and motivated by recent achievements in malaria control,3,4 the six GMS nations have endorsed a regional malaria elimination plan with an ultimate goal of eliminating Plasmodium falciparum malaria by 2025 and all malaria by 2030 in all countries of the GMS.5 Recently, after 3 years with no indigenous malaria cases, China was certified as malaria-free by WHO, marking a major success in the decades-long fight against this disease. However, various setbacks have been encountered in other GMS countries due to existing and emerging challenges (detailed in the following).
Malaria control and elimination rely on accurate and timely knowledge of the distribution of malaria incidence and prevalence, delivery of effective chemotherapy, and implementation of operative vector-management strategies. The complex and fast-evolving malaria epidemiology in the GMS is reflected in its immense spatial heterogeneity and the emerging dominance of Plasmodium vivax, a parasite species with remarkable resilience to conventional malaria control methods.6 In addition, artemisinin (ART) resistance in P. falciparum, detected initially in Western Cambodia a decade ago, has received augmented local and international concerns.7–10 Failure to contain ART-resistant parasites and the emergence of resistance elsewhere in the GMS escalated the urgency for a regional plan of malaria elimination.11,12 Further, the effectiveness of two core vector control interventions (insecticide-treated nets and indoor residue spraying) has been declining due to the development of insecticide resistance and increased outdoor biting of vectors.13,14 To address these problems, the Southeast Asia International Center of Excellence for Malaria Research (ICEMR) has developed a multidisciplinary program, aiming to understand how human migration, antimalarial drug resistance, and vector adaptations contribute to continuous malaria transmission at international borders so that integrative control strategies can be developed. To realize this scientific goal, we have strategically selected representative sentinel sites along the international borders of China, Myanmar, and Thailand, where malaria epidemiology is drastically different from each other. Using systems approaches and innovative technologies, we want to dissect the tripartite interactions among migrant human populations, diverse mosquito vectors, and multidrug-resistant (MDR) parasites to develop novel control strategies to propel the course of regional malaria elimination.
EPIDEMIOLOGY OF BORDER MALARIA
The distribution of malaria in the GMS exhibits extreme heterogeneity at both macro and microgeographical scales.15,16 The six GMS countries have advanced to different stages of malaria elimination, with Myanmar having the highest malaria incidence (almost 70% of the regional burden). Although border malaria (concentrated malaria transmission along international borders) is a shared phenomenon of each country,17 intensified control efforts have led to isolated pockets of malaria transmission.18 In Thailand, malaria has declined over the last several decades, but pockets of malaria transmission persist along the Thai–Myanmar border (Figure 1). Of the 927 border districts, 637 (69%) reported malaria incidence in the past 3 years and 307 (33%) in 2021. Similarly, during the final phase of malaria elimination in China, malaria in the border counties of Yunnan province displayed large spatiotemporal changes with incidence clustered in several hotspot townships.16 While the P. falciparum clusters shifted locations and cluster size each year, high-incidence vivax malaria clusters persisted.16 Within villages, malaria also exhibited evident transmission hotspots, probably depending on the local ecology of vectors.19,20
Another conspicuous change in malaria epidemiology is the increasing dominance of P. vivax malaria.21,22 Surveillance of clinical malaria cases at the China–Myanmar border detected an increase in the proportion of vivax malaria from ∼60% in 2011 to > 97% in 2016, with occasional vivax malaria outbreaks.21 Such a trend has persisted in more recent years (Figure 2). The proportional increase of vivax malaria is partially attributed to its ability to relapse, which requires 14-day primaquine (PQ) radical cure, a regimen with ubiquitously poor compliance. In a cohort of 7,000 village residents on the Western Thai border, we detected 410 malaria cases by microscopy in 6.5 years. Among them, 67 people had multiple malaria episodes within 1 year of the initial infection, and 60% of these recurring infections were due to P.vivax.23 The resilience of vivax malaria to conventional malaria control measures necessitates new tools for its elimination.
Risk factors for malaria transmission.
Risk factors for malaria infection vary by parasite species, geography, and demographic attributes. Plasmodium falciparum is more geographically restricted and clusters in rural, remote areas with poor healthcare access—especially along borders. In much of the GMS, P. falciparum infection clusters in adult males who are exposed to the parasite through travel to hotspots of the disease (e.g., forested areas).21,24,25 Certain occupations (e.g., farming, military) bear a significantly higher risk of malaria while students have an increased risk of vivax malaria.21,24 Individuals with poor access to health services, with linguistic barriers, of ethnic minorities, and without citizenship may also have a higher likelihood of infection.20,21,26 For both vivax and falciparum infections, individuals who have a history of malaria infection are more likely to have subsequent infections.21,24 Housing characteristics are also related to the risk of infection, presumably associated with Anopheles permeability (open structures, building materials, distance to mosquito breeding habitats, etc.).20,27,28 In addition, housing can be a proxy for other factors like socioeconomic status, which influences occupation and access to healthcare. Identifying high-risk populations facilitates the implementation of targeted malaria control measures. Delivering health education messages to hotspot villages29 and malaria prevention packages to forest-goers and farmers staying in farm huts will help change risky behaviors and reduce malaria infection.30,31
Malaria parasite detection and surveillance.
In low-endemic malaria settings in border communities, most Plasmodium infections appear to be asymptomatic and submicroscopic,32 requiring sensitive molecular tools for detection. We have demonstrated that submicroscopic infections can infect mosquitoes,33 constituting a critical reservoir for persistent transmission. In clinical settings, malaria diagnosis is routinely performed using light microscopy and rapid diagnostic tests (RDTs). RDTs have recently gained considerable traction in the GMS and play an indispensable role in evidence-based treatment, especially in hard-to-reach remote communities along international borders, where quality microscopy is often inaccessible. As most of the RDTs deployed in the GMS for P. falciparum are based on the detection of histidine-rich protein (HRP) 2 protein, our recent findings on the emergence of parasites with pfhrp2 deletion in the Western GMS suggest potential challenges for the continued use of such RDTs.34,35 Consistent with the suboptimal performance of RDTs against nonfalciparum and nonvivax human parasite species found in Southeast Asia,36 we also demonstrated the failure of a conventional RDT to diagnose high-density (> 500 parasites/mL) acute febrile infections of Plasmodium malariae and Plasmodium ovale in the China–Myanmar border area.37
Microscopy and RDTs have limited utility for active surveillance because parasite densities in asymptomatic infections are often below their detection thresholds. Recognizing these limitations, we have conducted studies to compare potential new solutions, aiming to identify pragmatic tools for disease surveillance in the GMS. The recent advent of an ultrasensitive RDT (uRDT) for P. falciparum, having a detection limit 10 times lower than conventional RDTs, prompted the team to investigate its utility for active surveillance.38 Our study conducted in endemic areas of Myanmar demonstrated that uRDTs have approximately 20% increased sensitivity in detecting subclinical P. falciparum infections when compared with standard RDTs.39 Ultrasensitive RDTs still have lower sensitivity than molecular assays and are unlikely to identify all subclinical infections, but they are a promising improvement in our ability to monitor P. falciparum. The increasing predominance of P. vivax demands the development of uRDTs for this species.
The program evaluated several molecular diagnostics, including qPCR, nested PCR to detect parasite rRNA genes, nested reverse-transcriptase PCR (nRT-PCR) to detect parasite rRNAs, and capture and ligation-probe PCR (CLIP-PCR) to detect parasite rRNAs in cross-sectional surveys.40,41 The rRNA-based method has the highest sensitivity and rivals that of high-volume PCR,42 but the RNA detection requires a much smaller blood volume and is more suitable for active surveillance in many places. Applying nRT-PCR to finger-prick blood samples from community surveys in Northeastern Myanmar uncovered an infection prevalence of nearly 20% compared with 1% by light microscopy,27,40 further demonstrating the feasibility and the gain of using a sensitive molecular tool. As costs are one major impediment to molecular testing, a simple and flexible method of sample pooling was devised,39 which can be tailored to different endemicities. As most infections in areas approaching elimination are asymptomatic and submicroscopic,33 molecular surveillance in sentinel sites is essential for guiding targeted control practices, determining the effects of control measures, and monitoring the progress toward elimination. Further fine-tuning these molecular tools to differentiate the drug resistant and sensitive parasites in a clinical setting would also be crucial for timely adjusting drug policies.
Migration and malaria introduction.
Border malaria poses a vital threat to malaria elimination and requires multinational cooperation.43 Heavy population flow along the extremely porous borders makes neighboring countries vulnerable to malaria introduction and reintroduction.44,45 Human migration may be partially responsible for the cross-national spread of ART-resistant strains with specific multidrug resistance genotypes.46 The association of a higher risk of malaria with the migrant population and those with travel to Myanmar highlights the significance of malaria introduction by migratory populations in the border region.20,47,48 Although passive case detection activity in the Southwestern border of China only showed strong evidence of imported P. falciparum malaria,47 subsequent genetic studies at the China–Myanmar border using microsatellite markers revealed genetically homogenous populations for both parasite species on both sides, indicating extensive parasite gene flow not constrained by the political border.49,50 Analysis of parasite migration patterns within and between the two sides of the international border detected unidirectional migration of parasites from Myanmar to China, providing genetic evidence of parasite migration in the border region. Especially for P. vivax, a parasite that can travel long distances by infected migrants as silent liver hypnozoites, there is an urgency to identify the sources and sinks of the parasites to enable timely targeted control. The use of polymorphic antigen markers such as Pvmsp3α and 3β has revealed highly diverse P. vivax populations in Western Thailand border despite low endemicity, and detected clonal expansion events in Southern Thailand, likely resulting from relaxed control efforts.51–53 Using microsatellite markers, we found drastically divergent P. vivax populations in the Eastern and Western Thailand borders, with the central malaria-free zone as a gene flow barrier.54 The possibility to distinguish these parasite populations using as few as four microsatellite markers will simplify the tracking of parasite migration, at least among the Thailand borders. We also found that microsatellites could be used to assess the temporal population changes as a means to monitor the progress of malaria control. Although the genetic diversity of P. vivax populations over time may remain high, the decreased multiplicity of infection and increased multilocus linkage disequilibrium may reflect a reduction in the parasite population size.55 In Eastern GMS, where P. vivax populations are less geographically isolated and genetically distinct, whole-genome sequencing (WGS) and the derived SNP barcode may be necessary to distinguish closely related parasite strains and identify the origins of the parasite.56,57 The genomic information from spatially representative parasite populations would identify potential migration patterns using shared identity-by-descent segments,56,58 providing the scientific basis for enhanced monitoring of parasite introduction by migrant populations.
Zoonotic Plasmodium knowlesi malaria.
Since the first cluster of Plasmodium knowlesi malaria cases in humans was reported in 2004 in Malaysian Borneo,59 reports of P. knowlesi incidence have increased strikingly, including in all countries of the GMS—Thailand,60–64 Laos,65,66 Cambodia,67 Myanmar,68,69 and Vietnam.65,70,71 This wide range of P. knowlesi in Southeast Asia largely reflects the distribution of the zoonotic hosts (the long-tailed and pig-tailed macaques) and vectors of the Leucosphyrus group of anopheline mosquitoes.72,73 This parasite is probably historically present in the GMS rather than newly emergent. In recent years, we and others have identified an increasing trend of clinical P. knowlesi cases in Thailand.63,64 Increased incidences of P. knowlesi are likely due to environmental changes such as deforestation, increased forest-related human activities, and potentially peridomestic transmission.74 Plasmodium knowlesi diagnosis is challenging75—it is often misdiagnosed by microscopy due to its resemblance to P. malariae and P. falciparum, current RDTs are not sufficiently sensitive to detect P. knowlesi, and confirmation requires the use of molecular methods.61,76 Its presence as coinfections with other human malaria parasites and in asymptomatic infections also complicates diagnosis and detection, resulting in an underestimate of its real burden.61,65,68,69,71,77,78 Since the regional malaria elimination efforts are meant to target all Plasmodium species,79 it is also time to consider eliminating P. knowlesi and other monkey malaria parasites infecting humans (P. cynomolgi, P. inui, etc.).80–82 The diverse factors associated with the transmission of these zoonotic malaria parasites present a challenge for their elimination, as conventional vector-based control efforts in the domestic environment are ineffective in protecting against sylvatic transmission. Strategies such as repellent and chemoprophylaxis targeting high-risk populations like forest-goers are advocated to accelerate malaria transmission in the GMS.31
MOSQUITO ECOLOGY AND INSECTICIDE RESISTANCE
Malaria vectors in the GMS consist of many Anopheles species with varying abundance and importance in malaria transmission among different geographical regions.83,84 Many vector species are in species complexes, including several morphologically similar species and possibly cryptic species. The abundance, diversity, distribution, survivorship, biting behaviors, and vectorial status of different vectors can be influenced by environmental changes, such as deforestation and extensive use of insecticides in both public health and agricultural sectors. As “forest malaria” is a major contributor to residual malaria incidence,85,86 deforestation and landscape changes will have a significant impact on vector ecology and malaria transmission.87 Our study conducted in the China–Myanmar border area showed that adult An. sinensis and An. minimus, the main malaria vectors in this region, had much higher survivorship in deforested than forested areas.88 Deforestation also enhanced the survival of An. minimus larvae and accelerated larval development.89 Our vector surveillance studies conducted in sentinel sites of China, Myanmar, and Thailand have detected major changes in Anopheles composition and seasonal dynamics (Table 1). An. minimus was the predominant vector in all the surveys.30,90–92 Consistent with An. minimus being a highly adaptive vector, population genetic analysis revealed similar population genetic structure of past and present An. minimus populations and substantial gene flow among different geographical populations.93 These studies also revealed increased abundance of other vectors such as An. annularis and An. barbirostris s.l., some of which may support outdoor transmission.30,90–92 The vector species composition is further complicated by the presence of morphologically identical cryptic species. In Western Thailand, An. minimus A and An. harrisoni (An. minimus C) are two cryptic species often found in the same locations.94 Our recent molecular studies of An. minimus species collected from Western Thailand showed that ∼11% of the morphologically identified An. minimus belonged to a cryptic species (lineage B), which deserves further investigation to understand its bionomics, vectorial status, and species evolution.93
Anopheles species compositions in different study sites and study periods at the international borders of the GMS
|Species||China–Myanmar border (2012–2014)a||Tak, Thailand (2011–2013)b||Tak, Thailand (2015)c|
Mosquitoes were collected by CDC light traps. This table illustrates major changes of primary vector species in different sentinel sites. Anopheles species with ≥ 1% abundance were listed by species names, while the rest was summarized as “Other Anopheles.”
a From Wang et al.,92 b from Sriwichai et al.,90 c from Sumruayphol et al.92
In addition to changes in vector species, the malaria vectors in the GMS showed different levels of adaptations to the microecology with dramatic variations among villages. Their different seasonal dynamics underlie their roles in malaria transmission in different seasons.90–92 Residual malaria transmission was traced to farm huts and outdoor agriculture sites, where human biting rates were the highest with An. minimus, An. dirus, and An. maculatus as the primary vectors.30 In Western Thailand, An. minimus and An. maculatus are the main vectors during the two annual malaria transmission peaks while An. minimus group is the key primary vector in the dry season,94 the Maculatus group is most abundant in the wet season with within-group species-specific variations.91 Collectively, this knowledge of the species composition, distribution, bionomics, and dynamics in the international border regions is needed to guide vector control efforts.
Extent, distribution, and mechanisms of insecticide resistance.
Fast emerging and increasing insecticide resistance of malaria vectors has been implicated as a significant threat to malaria prevention by vector control. Understanding the status, distribution, and mechanisms of insecticide resistance in local malaria vector populations is critical for resistance management and effective malaria control and elimination. We have been monitoring the resistance of malaria vectors to multiple insecticides using the WHO tube test in multiple study sites in China, Thailand, and Myanmar since 2011. The two best-known resistance mechanisms (target site resistance and metabolic detoxification) were investigated in field populations of Anopheles mosquitoes. High-level resistance to the four major classes of insecticides (pyrethroids, organochlorines, organophosphates, and carbamates) was observed in An. sinenesis populations from Southern and Central China,95–97 and the Eastern Coastal region of China.98 Three nonsynonymous knockdown resistance (kdr) mutations (L1014F, L1014C, and L1014S) were detected at codon L1014 of the para-type sodium channel gene in An. sinensis from China, and these kdr mutation alleles exhibited a patchy distribution in frequency from Southern to Central China. Near fixation of kdr mutation was detected in populations from Central China but no kdr mutations were found in Southwestern China, suggesting that kdr alone is insufficient to predict pyrethroid resistance.99 The G119S mutation of the ace-1 gene in An. sinensis was moderately frequent in Southern and Central China but fixed in the Eastern Coastal region of China.96–98 Recently, high-level resistance to deltamethrin (mortality rate, 40–80%) was observed in multiple Anopheles species, including An. minimus s.l. from Thailand in 2018, and the two major vector species complexes (An. hyrcanus s.l. and An. barbirostris s.l.) from Myanmar in 2019 (unpublished data). However, the kdr L1014 mutations or the ace-1 G119S mutation were not detected in any of the Anopheles species analyzed from Thailand and Myanmar, suggesting other mechanisms responsible for pyrethroid and organophosphate resistance (unpublished data). The classification and statistical regression analysis found that metabolic detoxification was the most important resistance mechanism, whereas target site insensitivity of L1014 kdr mutation played a less critical role.96 We have used transcriptome and WGS to identify transcripts and SNPs associated with insecticide resistance.100,101 These studies highlight the complex network of mechanisms conferring resistance to multiple chemical insecticides in mosquito vectors, and it has important implications for designing and implementing improved vector resistance management strategies.
ANTIMALARIAL DRUG RESISTANCE
ART-based combination therapies (ACTs) are the frontline treatment of P. falciparum and are also recommended as a unified treatment of P. vivax. The emergence of P. falciparum parasites resistant to ART and partner drugs significantly compromised the efficacies of two ACTs—artesunate-mefloquine (AS-MQ) and dihydroartemisinin-piperaquine (DHA-PPQ).102–107 Clinical ART resistance is manifested as delayed parasite clearance with a parasite clearance half-life of > 5.5 hours, compared with ∼2 hours typically associated with ART-sensitive parasites.8,108–110 Day-3 blood smear parasite-positivity is also a crude measure of ART resistance, with a 10% cutoff for suspected ART resistance.111,112 In vitro, ART resistance is measured by the ring-stage survival assay (RSA), which measures the survival rate of early ring-stage parasites exposed to a 6-hour pulse of 700 nM of DHA, with an RSA value of ≥ 1% considered as an indication of ART resistance.113 Genetically, mutations in the propeller domain of the Kelch-domain protein K13 were identified as the key determinants of ART resistance.114 Of the > 200 PfK13 mutations identified in the global parasite populations,115,116 many have been confirmed in clinical efficacy studies111,117,118 while some have been validated genetically for in vitro ART resistance.119–122
Monitoring clinical efficacy of ACTs in Western GMS.
We have focused our efforts on monitoring the emergence and spread of ART resistance in Myanmar, given its disproportionate malaria burden in the GMS and its bridging position with South Asia. In Northeastern Myanmar bordering China, the evaluation of DHA-PPQ in 71 patients with uncomplicated falciparum malaria in 2012–2013 demonstrated a 42-day cure rate of 100% and a day-3 parasite-positive rate of 7%.123 Similarly, we also found a 28-day cure rate of 100% for artemether-lumefantrine in 41 falciparum patients at the Western border of Myanmar in 2015, although the day-3 positivity rate exceeded 10% in the latter study.124 Assessment of 44 culture-adapted clinical isolates for RSA demonstrated increased ring survival rates in parasites with PfK13 mutations.125 In addition, day-3 parasite-positive isolates had ∼10 times higher RSA values than day-3 parasite-negative isolates. These studies set the stage for using in vivo efficacy study, in vitro RSA, and molecular surveillance as complementary approaches to monitoring ART resistance.
Longitudinal in vitro drug susceptibility and molecular markers of resistance.
Our efforts over the past decade to procure clinical isolates from the China–Myanmar border area and establish continuous culture have allowed us to follow the dynamics of in vitro drug susceptibility longitudinally.126–129 From these studies, in vitro sensitivities to 4-aminoquinolines, antifolates, and ARTs deserve some attention. Although chloroquine (CQ) has been withdrawn from treating P. falciparum malaria for some time, CQ resistance is consistently high, corresponding with the prevailing occurrence of the Dd2-like pfcrt genotype, the primary determinant of CQ resistance.126,128 The use of CQ as the frontline treatment of P. vivax malaria may have continually exerted collateral selection pressure on the sympatric P. falciparum. Similarly, although the antifolate drugs were withdrawn quite some time ago, parasites exhibited continuous increases in resistance to pyrimethamine, and major mutations in the pfdhfr and pfdhps genes mediating antifolate resistance remain highly prevalent.126,128 The drug that replaced CQ in this region is PPQ monotherapy,130 also the partner drug for the commonly used ACT, DHA-PPQ. Despite previous reports of clinical resistance to PPQ and identification of pfcrt mutations, which may be associated with PPQ resistance,131,132 recent studies showed that the efficacy of DHA-PPQ for uncomplicated P. falciparum malaria remained high.123,133 Parasites collected over the years were relatively susceptible to PPQ with temporal fluctuations in IC50 or IC90.129 We did not observe parasites with either plasmepsin 2/3 amplification or new pfcrt mutations (H97Y, F145I, M343L, and G353V),129 which were described in the DHA-PPQ-resistant populations in Cambodia.134–138
PfK13-mediated and non-PfK13 ART resistance mechanisms.
PfK13 mutations have also experienced drastic spatiotemporal changes in the GMS. In the Eastern GMS, the C580Y mutation was predominant and has swept rapidly across Cambodia and the Eastern GMS.114,115,122 In the Western GMS, the F446I mutation is the most prevalent.139–141 Table 2 summarizes the results from our molecular surveillance of PfK13 mutations in the Western GMS. An updated distribution map of major PfK13 mutations in endemic sites of the GMS is shown in Figure 3. On the Eastern border of Myanmar, F446I has gained a steady increase in prevalence between 2007 and 2013.127,140 In the 2014–2016 samples, the G533S mutation emerged and became the second most prevalent at 44%. This new mutation was associated with increased RSA values.127 Analysis of asymptomatic P. falciparum infections from cross-sectional surveys conducted in the Eastern, Northern, and Western border areas of Myanmar during 2015–2018 detected the F446I mutation only on the Eastern border, suggesting that ART resistance has not spread to or emerged in the Western and Northern borders (Table 2).142 To determine whether the PfK13 mutations found in the Western GMS indeed confer ART resistance in vitro, we engineered the F446I, N458Y, C469Y, F495L, and C580Y mutations in the 3D7 background and confirmed that the N458Y and C580Y mutations conferred significant increases in ring survival rate.121 Conversely, reverting the F446I, N458Y, C469Y, and C580Y mutations to the wild type (WT) in field isolates resulted in significant decreases in RSA values in all except for the C469Y mutation. Although all tested PfK13 mutations incurred different levels of fitness cost in the transgenic parasites, the F446I and C580Y mutations were almost as fit as the WT,121 which may explain their high prevalences in the field parasite populations. In addition, transgenic parasites with these two mutations also exhibited a prolonged ring stage, presumably enabling the parasites to better survive ART treatment, which has a short half-life.
Amino acid substitutions detected in the PfK13 gene of P. falciparum populations at different border areas of Myanmar
|Mutation||China–Myanmar border (and East Myanmar)||Banmauk, North Myanmar||Paletwa, West Myanmar|
|2007–2012 (N = 191)a||(2013–2016) (N = 74)b||2017–2018 (N = 53)c||2015 (N = 30)d||2017 (N = 22)c|
|K189T||3 (1.6)||2 (2.7)||4 (17.4)||3 (10.0)||–|
|F446I||52 (27.2)||44 (59.5)||–||–||–|
|N458Y||1 (0.5)||1 (1.4)||–||–||–|