Background
The emergence of artemisinin-resistant
Plasmodium falciparum within the Greater Mekong Subregion (GMS) has led to the region setting elimination targets for all human malaria by 2030 [
1]. In working towards this goal, the incidence of malaria cases and deaths in the GMS fell substantially by 75% and 93%, respectively, between 2012 and 2017 [
2]. Monitoring and surveillance are critical to the elimination of malaria ensuring that progress towards malaria elimination targets can be accurately tracked and ultimately accelerated. However, as the region transitions towards malaria elimination, surveillance becomes increasingly difficult because malaria becomes concentrated in discrete geographical foci, such as border and hard-to-reach areas, and in high-risk populations such as migrant workers and residents of highly forested areas [
3‐
5]. To capture these infections, in many remote areas, surveillance and malaria control strategies are dependent on passive case detection (PCD) and active case detection (ACD) provided by a village health volunteer (VHV) network who administer malaria testing by rapid diagnostic test (RDT) as well as treatment. However, subclinical
Plasmodium spp. infections often go undetected because asymptomatic individuals are less likely to seek testing and treatment and, importantly, are generally below the detection limit of conventional RDT diagnostics used in the field. Undetected, and therefore untreated,
Plasmodium spp. infections may be an important source of residual malaria transmission [
6‐
10], and failure to detect and eliminate all infections may hinder malaria elimination targets
.
Integration of more sensitive diagnostic measures into the VHV network may improve surveillance of residual malaria transmission in hard-to-reach areas in the region. Molecular and serological assays can determine residual malaria transmission not detected by routine diagnostics in the field such as RDT and microscopy. Microscopy is estimated to miss approximately 50% of infections when compared to molecular methods such as polymerase chain reaction (PCR), and the proportion of missed infections may be greater than 80% in areas of low transmission (defined at PCR prevalence <10%) [
11]. However, the application of sensitive molecular methods such as PCR to detect malaria is mostly utilised as a research tool in many malaria-endemic settings and is yet to be approved and incorporated into routine surveillance in the GMS. Similarly, measuring antibodies specific for malarial antigens is not approved for routine use and may also be a useful surveillance tool to monitoring ongoing malaria transmission in regions approaching malaria elimination as it has the potential to measure both current and recent malaria exposure [
12]. Until point-of-contact molecular and serological surveillance tools for malaria become more widely available and approved for use in national malaria control programmes, centralised use of these approaches will be necessary. However, sample collection for surveillance activities may be implemented at the village level by VHV. To date, few studies have been performed in Southeast Asia to investigate the utility of molecular and serological surveillance, and none has been incorporated into the VHV network. Serial cross-sectional research surveys have demonstrated higher blood-stage antimalarial IgG levels and seropositivity amongst those with PCR-detectable subclinical
P. falciparum and
P. vivax infection compared to uninfected individuals [
13‐
16], and geospatial analysis has shown that antimalarial antibodies are predictive of ongoing malaria transmission [
17]. While these studies suggest the use of serological surveillance may be appropriate for the detection of residual malaria transmission, the feasibility of integrating the approach and the data on residual malaria transmission that would be captured into the PCD/ACD VHV-led testing and treatment surveillance network in the GMS is unknown. To address this knowledge gap, we integrated the collection of participant samples for molecular and serological surveillance into VHV-delivered community-based malaria programmes in Southeast Myanmar to understand the surveillance data that can be captured at this level to inform surveillance of malaria and targeted interventions in elimination settings in the GMS.
Discussion
VHVs are the cornerstone of malaria control and surveillance activities in many malaria-endemic settings. By training VHV delivering routine malaria services to integrate sample collection for molecular and serological surveillance in their village, we demonstrated that surveillance using highly sensitive molecular methods in samples collected by VHV is feasible and can detect a significant subclinical reservoir of infection undetected by standard surveillance using RDTs. Furthermore, serological data reflected monthly trends in P. falciparum and P. vivax infection, even in those with no detectable parasites, indicating that this surveillance approach can capture significant levels of ongoing undetected seasonal malaria transmission. Integration of molecular and serological surveillance into the extensive VHV network may be an effective tool for monitoring residual malaria transmission in hard-to-reach pre-elimination settings by supplementing an established VHV surveillance network currently dependant on RDTs. Further, the addition of this kind of supplemented surveillance could support centralised malaria control strategies such as resource allocation until more suitable field-deployable tests are available to provide VHV with real-time treatment decision-making tools. Improving malaria surveillance will advance progress, and tracking, towards malaria elimination goals in the GMS.
RDT diagnosis and reporting of malaria cases by VHVs is the key indicator for progressing towards malaria elimination goals in the GMS. Like in much of the region, a large proportion of
Plasmodium spp. infections found in participating villages were missed by RDT diagnosis with ~22-fold greater prevalence of qPCR-detectable infections over the entire study period. There was significant variation in qPCR-detectable infection over time with peaks of
Plasmodium spp. infection in the rainy season, where malaria transmission increases, which were not detected by RDT, nor subsequently treated. There is increasing evidence that these untreated molecular-detected infections have the potential to contribute to ongoing malaria transmission, and despite not being transmitted as effectively to mosquitoes as microscopically detected infections, the greater prevalence of subclinical compared to clinical infections in low-transmission settings is cause for concern [
25]. Detection of these infections is becoming increasingly difficult because in the elimination phase, malaria is concentrated in hard-to-reach areas and high-risk populations. Therefore, additional surveillance tools deployed at the VHV level may help target interventions aimed at reducing the subclinical burden of malaria and advance the malaria elimination agenda in this region.
Serosurveillance does not provide a diagnosis of a current
Plasmodium spp. infection. Rather, at a population level, antimalarial antibodies may indicate recent
Plasmodium spp. transmission and exposure and therefore represent a useful tool for malaria microstratification to inform resource allocation and targeted implementation of malaria control and elimination interventions by NMCPs if integrated into national surveillance. The observed peaks in anti-
P. falciparum merozoite IgG seroprevalence coinciding approximately with peaks of PCR-detectable
P. falciparum infection, even in parasite negative individuals, indicate that malarial serology captured ongoing malaria transmission in the population not detected by routine RDTs. However, the magnitude of antibody seroprevalence was significantly greater than the prevalence of PCR detectable infections. Caution must be exercised in estimating the prevalence of circulating infections based on serology, which may be overestimated particularly in regions where the malaria burden is transitioning from malaria control to elimination phases, as antibodies can persist for extended periods [
26,
27]. This study was conducted amid Myanmar’s transition from malaria control to elimination and following large-scale reductions in
Plasmodium spp. transmission in the study region [
28]. The higher seroprevalence compared to PCR prevalence of
P. falciparum and
P. vivax infections may reflect cumulative incidence in the region over this time period [
28,
29]. Between 2005 and 2014, immediately prior to the commencement of this study in 2015, the prevalence of malaria detected by RDT declined in Kayah, Kayin and Bago states by more than 74% (from 3.2 to 1.7%, 0.86 to 0.52%, and 0.44 to 0.07%, respectively) [
29]. The historical prevalence in the 10 years prior to this study are considerably higher than the 0.16% RDT positivity reported here and may result in the cumulative incidence of antimalarial antibody responses observed because the mean antimalarial antibody half-lives of the antigens included in this study (selected because their immunogenicity and half-lives are relatively well-characterised in the GMS) may range between 6 months and 7 years IgG [
27,
30,
31]. As such, serosurveillance may be a more informative tool as antibodies wane in the broader population and is limited to only exposed populations in elimination phases or alternatively different sets of antigens could be defined which can quantitatively capture very recent molecular infection events [
14]. This study was not designed to identify serological biomarkers of recent exposure to molecularly detectable infections as this would require detailed longitudinal follow-up and accurate participant exposure history. Several recent studies have demonstrated the utility in combining novel combinations of antigens to predict recent
Plasmodium spp. infection exposure detected by microscopy, clinical incidence and relapse with
P. vivax, across diverse malaria-endemic populations [
14,
32‐
34]. However, the efforts to identify reliable, quantitative antibody signatures of recent exposure to malaria parasites are ongoing [
35,
36], and there is currently a lack of consensus on antigen-specific antibody responses which discriminate against recent and historical exposures, in particular, of molecularly detectable infections. We investigated
P. falciparum antigens based on the 3D7 allelic variant. There is limited data on antigen allelic diversity in Myanmar, but previous studies have shown that the majority of circulating MSP alleles are of the 3D7 type [
37] and that variation in IgG response to the different allelic variants of
PfMSP2 across multiple sites in Myanmar and the Greater Mekong Subregion is minimal [
38]. Furthermore, antibody responses specific for epitopes within conserved domains of merozoite antigens have been shown to illicit strain transcending antibody responses [
39], so inclusion of additional antigenic variants may not have impacted our study conclusions. However, further refinement of choice and number of antigens and allelic types across different areas of Myanmar and the GMS is warranted to inform the utility of serological findings in the region. Regardless of ultimate antigen selection or malaria burden phase, any new tools need to be validated with respect to the health system platform, such as the VHV-led delivery of malaria services, which will ultimately be integrated into.
In individual-level analyses, we observed a reduction in the odds of concurrent qPCR detectable
P. falciparum infection in participants seropositive for
P. falciparum-specific IgG. Recent longitudinal data from the GMS has shown that within infected individuals, parasite density oscillates frequently and that spontaneous clearance of PCR-detectable
P. falciparum infection (i.e. in the absence of treatment) occurs regularly with the median time to resolution of infection of around 2 months for
P. falciparum and 6 months for
P. vivax [
40]. We hypothesise that boosting of antibodies upon exposure to parasites may then contribute to clearance of subclinical parasitaemia which may explain the observed reduction in the odds of a contemporaneous qPCR-detectable
P. falciparum infection in seropositive individuals when sampled in the VHV network. This study provides evidence of potential protective immunity against molecular-detectable infections, which to date has only been established to reduce clinical malaria and high parasitaemia in numerous high-transmission populations [
41]. We were unable to show protection against prospective infection; however, there were fewer repeated measures (returning participants were less than 20% of all samples) with large variation in the time between tests. Further studies investigating the response to and role of anti-malarial antibodies in the persistence of molecular-detectable subclinical infections would be valuable to inform the utility and sampling framework of serological surveillance in areas where the majority of the malaria burden remains undetected by conventional surveillance.
The aim of this study was to investigate malaria incidence and serological data within an existing national malaria surveillance strategy, in this case, the VHV network. In the national VHV testing strategy, some individuals will be sampled multiple times and others only once, that is, malaria incidence is representative of the population that is sampled by VHV which may not be truly representative of the general population per se. Repeated sampling of a subset of individuals within the same population risks overestimating malaria incidence when the same individuals present multiple times with a single infection. In this study, which accounted for repeated measures in statistical analyses, only one participant contributing multiple samples was qPCR positive more than once, in samples collected 10 months apart. Given the estimated clearance time of sub-clinical
P. falciparum infection in the region is estimated to be under 3 months [
40], these likely represent two separate infection events. Therefore, in the present study, multiple samples from individual participants would not have biassed estimates of malaria incidence in samples collected by VHV. Nonetheless, accurate estimates of malaria incidence due to repeated sampling of a subset of individuals should be considered in the broader national malaria surveillance strategy, given that the sampling strategy is the same as current RDT testing implemented nationally at the VHV level.
Sampling at the VHV level may facilitate increased surveillance in rural and hard-to-reach areas. In this study, VHVs aimed to collect a minimum of 20 RDTs each month in accordance with programme implementation through PCD/ACD; however, data could not be stratified based on the VHV case detection method as this data was not collected. As such, we are unable to determine whether PCD alone would be sufficient to accurately capture the subclinical malaria burden or changes in serological markers reported here. The total number of RDTs declined across the study period, and this is most likely due to the declining motivation of the VHV to undertake RDTs which has been described in the region in areas approaching elimination where positive RDT results are infrequent [
42]. This was not reflected in sample collection, the total number of which fluctuated across the study. Participation in sample collection, however, occurred in less than half of participants receiving testing by RDT which, from anecdotal reports, may have been due to reluctance to commit further time to complete the additional informed consent and testing procedures (in Myanmar, dried blood spot sampling is not part of routine VHV services and considered research; therefore, additional consent is required). However, participants contributing samples were broadly similar to those contributing RDTs. Until dried blood spot sampling is considered part of routine sample collection, routinely collected RDTs could be utilised as an alternative source of sample for molecular and serological surveillance of malaria to overcome participation issues [
43‐
45], but their utility for molecular assays, given low parasite densities, needs to be investigated in the Myanmar context. Importantly, capacity building for local laboratories needs to be prioritised as well as understanding how this data will operationally be integrated into the national surveillance electronic system in order to inform programmatic decisions.
Acknowledgements
We would like to thank the local communities and VHVs for their participation in the research; KMSS staff Win Tun Kyi, Augustine Tual Sian Piang, Min Thant Zin Latt, Nwe Ni Aye and Khine Zar Lwin (KMSS National Office, Yangon, Myanmar); Joseph Maung Win, Tun Tun Aung, Richard Joseph, Shein Thu Ag, Thet Naing and Myo Tint (KMSS Yangon, Yangon, Myanmar); Ludovico Saw Piko, Benedetta, Win Win Aye, Ngo Petru, Maurice Nyo, Daniel Win and Saw Golbert (KMSS Taunggo, Taunggo, Myanmar); Albino Htwe Win, Tin Aung, Perpetua Aye Aye Mon, Mupaula, John Bosco, Alfred, John Min Aung and Poe Rah (KMSS Loikaw, Loikaw, Myanmar); Paul Thar San, Saw Isidore, Theresa, Saw Micheal and Saw Pho Muo (KMSS Hpa An, Hpa An, Myanmar) for the local advocacy, coordination and preliminary planning of field work; Burnet Institute staff Toe Than Tun, Poe Poe Aung, Phone Myint Win, Poe Poe Aung, Nyi Nyi Zaw, Htin Kyaw Thu, Ai Pao Yawn, Lia Burns, Naanki Pasricha, Ricardo Ataíde and Nicole Romero for contributing towards technical, coordination and management support and contextual inputs; Alissa Robertson and Josh Charles for the assistance in sample processing; and Myanmar Ministry of Health and Sports staff Dr. Thandar Lwin, Dr. Thaung Hlaing, Dr. Than Naing Soe and Dr. Kyawt Mon Win for the technical and administrative support.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.