Program for Resistance, Immunology, Surveillance, and Modeling of Malaria in Uganda
Lead Institution: University of California, San Francisco
The Program for Resistance, Immunology, Surveillance, and Modeling of Malaria in Uganda (PRISM) represents the East African region for the ICEMR network. Uganda is emblematic of the challenges faced by high burden countries, where routine surveillance systems are inadequate to assess trends in the burden of malaria or to monitor the impact of control interventions. Through PRISM, researchers have implemented a comprehensive malaria surveillance program including enhanced health facility-based surveillance and detailed longitudinal cohort studies in areas with differing transmission intensities. Complementary laboratory-based studies include surveillance for markers of antimalarial drug and insecticide resistance, serologic measures of malaria exposure, highly sensitive molecular assays for the detection of asexual and sexual parasites, and membrane feeding assays to assess human to mosquito transmission. These studies have greatly improved our understanding of the epidemiology of malaria in Uganda and of the impact of control interventions. In recent years, the program has expanded its malaria surveillance work and the scope of longitudinal studies to address more fundamental questions about interactions between the parasite, mosquito vector, and human host.
The program consists of five research projects:
- Surveillance project: individual patient level data is collected at 70 health facilities located across Uganda. These data are used to monitor temporal and geographic trends in malaria burden and evaluate the impact of population level control interventions.
- Resistance project: uses samples collected over time at multiple sites to characterize the evolution of phenotypic and genotypic markers of drug and insecticide resistance and assess the impacts of these markers on malaria transmission.
- Epidemiology project: uses longitudinal samples from cohorts to characterize factors that determine whether sporozoite inoculation (bite of an infected mosquito) results in the establishment of blood stage infection and characterize factors affecting the duration, density, and clinical consequences of blood stage infections.
- Transmission project: uses longitudinal samples from cohorts to determine factors associated with gametocyte production (the sexual stage of the parasite life cycle in the human host, which infects mosquitos), evaluate infectivity of the human host to mosquito vectors, and characterize the human infectious reservoir.
- Implementation project: using health facility-based malaria surveillance data, the incidence of malaria in target areas around the health facilities is being estimated. These data and complementary cross-sectional surveys are being conducted as part of a cluster randomized trial of two different newer generation long-lasting insecticidal nets (LLINs) in the context of a national LLIN distribution program.
These highly interrelated projects are being conducted in settings with varied malaria epidemiology and differing population level control intervention to provide critical information needed to optimize strategies for the control and ultimate elimination of malaria in Uganda.
The use of enhanced health-facility based surveillance to monitor temporal and geographic trends in malaria burden and assess the impact of population level control interventions
Malaria surveillance, which encompasses monitoring and evaluation of malaria control efforts, is essential to guide program planning and management. Most malaria control programs rely on routinely collected health facility-based data for surveillance needs. However, these data are often inadequate to monitor trends due to incomplete reporting, poor accuracy, limited diagnostic testing, and the reliance on aggregate measure of disease burden. In 2006, the Uganda Malaria Surveillance Program (UMSP) was established to collect high quality malaria surveillance data at six high volume public health facilities, also known as Malaria Reference Centers (MRCs), in collaboration with the Ugandan National Malaria Control Division. With support from PRISM, the surveillance network was expanded from 2014 to 2020 to include 70 MRCs located in 38 districts across the country. Individual level data is collected electronically from all outpatients that present to these MRCs. These data have been used to monitor temporal and geographic trends in malaria burden and assess the impact of population level control intervention. For example, in the setting of repeated national LLIN distribution campaigns PRISM researchers showed that stopping IRS resulted in a 5-fold increase in malaria incidence within 10 months, but reinstating IRS led to an over 5-fold decrease within 8 months. In areas where IRS was initiated, malaria incidence dropped by 85% after 4 years of sustained use. These data highlight the benefits of adding IRS for sustained period in a country when LLIN coverage is high and the risk of malaria resurgence if IRS is stopped.
Since 2018 an emphasis has been placed on collecting data on village of residence for all patients presenting to the MRCs. These data along with the identification of target areas around the MRCs and enumeration surveys to estimate the populations of these target areas have allowed researchers to generate estimates of malaria incidence within the target areas. These estimates of malaria incidence will be the primary outcome of a cluster randomized trial that has been embedded into a universal LLIN distribution campaign that was conducted throughout Uganda from 2020-2021. In this cluster randomized trial, PRISM will be comparing two newer generation LLINs: PermaNet 3.0 LLINs treated with a pyrethroid plus piperonyl butoxide (PBO) and Royal Guard LLINs treated with a pyrethroid insecticide plus pyriproxyfen.
The use of detailed cohort studies to improve our understanding of the epidemiology of malaria in different settings
PRISM has been at the forefront of efforts to conduct detailed longitudinal studies in representative cohorts in a variety of epidemiological settings. These studies involve enrolling residents of households representative of populations living in areas with differing transmission intensity and community level control interventions. The first series of cohort studies (PRISM 1 cohort studies) were conducted in three areas with different transmission intensity from October 2011 through June 2016.
The PRISM 1 studies found a dramatic decrease in the burden of malaria at the highest transmission intensity site (Nagongera sub-county, Tororo District) following the implementation of indoor residual spraying (IRS) in the setting of universal LLIN distribution. To further investigate this phenomenon and better understand the epidemiology of malaria in the setting of successful malaria control intervention, a PRISM 2 cohort study was conducted in Nagongera from October 2017 through October 2019. Methods similar to the previous cohort studies were used, with more frequent routine visits (every four weeks). Over the two years of follow-up, only 38 episodes of malaria were diagnosed (incidence 0.04 episodes/person/year) and there were no cases of severe malaria or malaria deaths. After five years of intensive vector control measures in Tororo, the burden of malaria was reduced to very low transmission levels. However, a significant proportion of the population remained parasitemic, primarily school-aged children with sub-microscopic parasitemia, providing a potential reservoir for malaria transmission.
Currently PRISM researchers are conducting a third cohort study (Border Cohort Study) in an area bordering one district (Busia) where IRS has not been implemented and the burden of malaria is high and in an adjacent district (Tororo) where IRS has been implemented for over 5 years and the burden of malaria is expected to be much lower. Similar methods are being used as in the PRISM 2 cohort study to allow for comparison of key malaria metrics such as malaria incidence, repeated measures of parasite prevalence, and a variety of entomological measures across a continuum of transmission intensity.
The use of molecular tools to better understand the epidemiology of blood stage infections
In highly endemic countries like Uganda, the malaria parasite reservoir is dynamic, and includes symptomatic and asymptomatic infections. Asymptomatic infections comprise a large proportion of the parasite reservoir and are important because studies have shown that even submicroscopic (low-density infections that are below the level of detection by microscopy) asymptomatic infections can transmit parasites to mosquitoes. In the PRISM cohort studies, researchers used a highly sensitive quantitative PCR assay, in addition to microscopy, to longitudinally assess parasite density in each participant. This allowed them to characterize the size and composition of the P. falciparum parasite reservoir over time in the PRISM 1 study and to examine the impact of indoor residual spraying of insecticide (IRS), a highly effective vector control technique, on the parasite reservoir. It was found that IRS was associated with a significant decline in the incidence of symptomatic malaria in all age groups (episodes per person per year declined from 3.98 to 0.13 in children aged <5 years, 2.30 to 0.15 in children aged 5–10 years, and 0.41 to 0 in adults). While IRS significantly reduced the prevalence of parasitemia, prevalence remained high (pre-IRS to post–third round: 58.5% to 11.3% in children aged <5 years, 73.3% to 23.7% in children aged 5–10 years, and 52.2% to 15.4% in adults). Notably, in all age groups, the majority of the residual parasite reservoir was submicroscopic. These study findings may help explain the rapidity of resurgence of malaria following the withdrawal of IRS observed in northern Uganda and elsewhere, and highlight that a combination of IRS with other effective interventions may be necessary as we aim to eliminate the infectious reservoir.
In addition to characterizing infections by qPCR, PRISM also use amplicon deep-sequencing to genotype parasite DNA. This allows them to follow individual clones (strains) in cohort participants and study the dynamics of these infections over time. Because polyclonal infections can occur due to co-infection (one mosquito bite transmitting multiple clones) or superinfection (multiple bites), genotyping to distinguish the clones present in an infection is necessary to assess the molecular force of infection (mFOI), or the incidence of genetically distinct parasite clones acquired over time, and to estimate the rate of clearance of infections. In the PRISM 2 cohort, amplicon sequencing was used to estimate the clearance of asymptomatic P. falciparum infections and found that females consistently cleared their infections at a faster rate than males, even after adjusting for age, baseline infection status, and parasite density. No evidence was found for a sex-based difference in exposure to infection through behaviors or as measured by mFOI. These findings have motivated collaborating immunologists to better characterize sex-based differences in the host response to the malaria parasite.
PRISM researchers have also used amplicon deep-sequencing to better understand transmission dynamics. By carefully characterizing infections in human participants, researchers can track which parasite clones are transmitted to mosquitoes during membrane feeding experiments. Moving forward, researchers will continue to use both quantitative PCR and amplicon deep-sequencing to elucidate the host dynamics of infections in the PRISM Border Cohort. PRISM researchers also plan to characterize the prevalence of non-P. falciparum species in the newest cohort.
The use of molecular tools and membrane feeding assays to better understand human to mosquito transmission
PRISM examined the human infectious reservoir for malaria in two distinct cohorts. Key questions were: the contribution to transmission of infections in relation to their detectability by different diagnostics (microscopy, molecular quantitative PCR); the contribution of different age groups to transmission; and the duration of parasite carriage in relation to their transmission potential.
In the first cohort (Tororo District; October 2017-October 2019), parasite prevalence by qPCR was low at enrolment (17.4%) and both parasite prevalence and parasite density among infected individuals declined during follow-up. Within this cohort, representative of many African settings where malaria transmission has declined considerably following intensive malaria control, researchers used mosquito feeding assays and molecular assays for gametocyte carriage and clonal complexity to examine the human infectious reservoir for malaria. They found that among parasite-positive individuals, the prevalence of the transmissible stage of malaria parasites (gametocytes) was dependent on clinical status at presentation. Among clinical malaria cases, many of whom had evidence of a recent incident infection, gametocyte prevalence was 28.9%. By comparison, gametocyte prevalence was 67.6% among asymptomatically individuals (p=0.033). They performed a total of 538 mosquito membrane feeding assays on blood samples taken from clinical malaria cases or asymptomatically infected individuals. At least one infected mosquito was observed in 39 (7.2%) of these experiments, with 446 (1.2%) of 37,404 mosquitoes becoming infected. Researchers observed that gametocyte density was strongly predictive of the likelihood of infecting at least one mosquito and the proportion of mosquitos that became infected. Gametocyte density distributions differed between populations and densities were overall highest in asymptomatic microscopy-detected malaria infections and lower in clinical malaria cases and asymptomatic submicroscopic infections. They also found indications that clinical symptoms reduce the transmissibility of gametocytes. When concurrently considering their prevalence in the population and their infectivity to mosquitoes, asymptomatic microscopy-positive individuals comprised 83.8% of the infectious reservoir, with asymptomatic qPCR-detected infections being responsible for 15.6% and symptomatic infections being 50 responsible for 0.6% of transmission. They further estimated that children aged 5–15 years were responsible for 58.7% of the infectious reservoir, followed by children younger than 5 years (25.8%), and those 16 years or older (15.6%). Researchers longitudinally monitored infections with repeated mosquito feeding assays in 75 individuals. Remarkably, four children (0.8% of the total population) were responsible for 62.6% of all infected mosquitos. Infections in these children were sometimes chronic but at other times of short infection duration.
PRISM researchers are currently performing a similar study in an area of markedly higher transmission intensity (Tororo-Busia border area). qPCR parasite prevalence at the start of the study (August-September 2020) was relatively high in the first 6 months of follow-up. In a total of 340 feeds, 33 infectious (9.7%) resulted in at least one infected mosquito with 1.2% of all mosquitos (283/23247) becoming infected. Despite markedly different transmission intensity, findings were markedly similar. Among 645 qPCR positive individuals who were randomly selected for gametocyte quantification, gametocyte prevalence was again lower in clinical cases and gametocyte density showed a very strong association with mosquito infection rates. Again, very few of the infectious individuals were individuals aged >15 years (1/33 infectious individuals; 1/283 infected mosquitos). The researchers conclude marked consistency between sites with a dominant role for asymptomatic infections in children in sustaining transmission to mosquitos.
Temporal and geographic trends in measures of antimalarial drug and insecticide resistance
Drug resistance. PRISM researchers have been monitoring trends in drug resistance using samples from a range of sources. They have conducted regular surveillance of known markers of drug resistance, with surveys at three sites in earlier years, 10 sites in 2016-17, and 16 sites around Uganda since 2018. Considering markers of resistance to older drugs, prevalences of pfcrt and pfmdr1 mutations associated with resistance to chloroquine and related aminoquinolines have decreased to low levels; prevalences of 5 mutations in pfdhfr and pfdhps that mediate moderate resistance to sulfadoxine-pyrimethamine (SP) remain very high across the country, and prevalence of two additional mutations, pfdhfr 164L and pfdhps 581G, have increased, especially in western Uganda. Use of intermittent preventive treatment in pregnancy with SP selected for mutations associated with resistance. Of greatest concern are mutations in pfK13 that may mediate resistance to artemisinins. They have seen increasing prevalence of three PfK13 mutations, 469Y and 675V in northern Uganda, and very recently 469F in southwestern Uganda. PRISM researchers also regularly characterize ex vivo drug susceptibility in P. falciparum isolates collected near their Tororo laboratory, and in June-July 2021 researchers extended these studies to include samples collected at a site in Agago District in northern Uganda. Studies to characterize associations between genotypes of interest and ex vivo phenotypes are underway. Overall, surveillance and laboratory studies in recent years have shown improved susceptibility to aminioquinoline antimalarials, signs of worsening resistance to SP, and worrisome emergence of a number of markers that predict resistance to artemisinin antimalarials. Continued surveillance and laboratory and clinical studies to assess drug resistance are a high priority.
Insecticide resistance. Routine monitoring of insecticide susceptibility to pyrethroids revealed high levels of resistance to class I (Permethrin) and class II (Deltamethrin) pyrethroids which are primarily used on long lasting insecticidal nets (LLINs). Insecticide resistance to another insecticide class namely: the carbamates (with bendiocarb) has also been documented. Examination of underlying mechanisms driving resistance this class revealed absence of target site mutation G119S in acetylcholinesterase-1 gene, widely found in West Africa, but a significant and unexpected association of salivary gland gene expression in bendiocarb survivors. Over the years, PRISM has continually examined several mosquito populations for novel markers of insecticide resistance and this effort paid off with the discovery of two novel single nucleotide polymorphisms (SNPs) namely; a cytochrome P450 Cyp4j5 and the esterase Coeae1d associated with metabolic resistance to pyrethroids. The PRISM project contributed samples to the Anopheles gambiae 1000 genomes project examining the genetic diversity of this African Malaria vector. This resource has been very instrumental in the discovery of additional modes of insecticide resistance including the importance of copy number variation in mediating insecticide resistance. PRISM has collaborated with other ICEMRs on mapping out insecticide resistance across the ICEMRS and closing knowledge gaps. They have also documented evidence of contemporary gene flow between sympatric Anopheles gambiae s.s and Anopheles arabiensis which is of concern in respect to the increased abundance of Anopheles arabiensis observed in response to indoor residual spraying campaigns. They are currently examining the implications to pyrethroid resistance in the recently discovered and rapidly spreading triple mutant found in a cluster of cytochrome P450 genes, constituting a nonsynonymous point mutation in Cyp6p4 (I236M), a Zanzibar-like transposable element (TE) and a duplication of the Cyp6aa1 gene in samples collected from 11 sites. This triple mutation is associated with high-level metabolic resistance in Anopheles gambiae. Overall, PRISM’s surveillance of insecticide resistance has shown high levels of resistance to pyrethroids and emerging resistance to carbamates. PRISM researchers keep monitoring for insecticide susceptibility to other classes especially organophosphates (Pirimiphos methyl) due to its application to indoor residual spraying. They intend to widen their scope of monitoring insecticide susceptibility to new generation insecticides such as clothianidin (a neonicotinoid) and clorfenapyr (a pyrrole) which have application in current vector control.
There has recently been a dramatic increase in the scale up of control interventions and reduction in the burden of malaria across Africa. However, progress has not been uniform, and in fact has been slowest in countries with the highest burden, such as Uganda. Malaria covers a wide range of epidemiological settings in the country. This ICEMR is conducting studies in health centers and cohorts around Uganda, ranging from areas of relatively low transmission intensity to areas with some of the highest transmission intensities recorded in the world. Researchers hope to use the varied settings to evaluate intervention strategies and assess optimal control methods.
View Associated sites for the East Africa ICEMR in a larger map
Map description: Sites associated with the East Africa ICEMR: Districts of Mubende, Kayunga, Kyegegwa, Kibaale, Hoima, Masindi, Kiryandongo, Arua, Koboko, Moyo, Nwoya, Amuru, Gulu, Omoro, Lamwo, Kitgum, Agago, Oyam, Kole, Apac, Kwania, Kaabong, Kapelebyong, Amuria, Kumi, Bukedea, Buyende, Kaliro, Luuka, Jinja, Mayuge, Busia, Kanungu, Amolatar, Dokolo, Alebtong, Otuke, Tororo.
Principal Investigator: Grant Dorsey, MD, PhD
- Surveillance: Joaniter I. Nankabirwa, Makerere University and Infectious Diseases Research Collaboration (IDRC), Uganda
- Epidemiology: Moses R. Kamya, Makerere University and Infectious Diseases Research Collaboration (IDRC), Uganda
- Resistance: Samuel Nsobya, Makerere University and IDRC, Uganda
- Transmission: Sarah Staedke, London School of Hygiene and Tropical Medicine, U.K.
- Implementation: Moses R. Kamya, Makerere University and Infectious Diseases Research Collaboration (IDRC), Uganda
- Infectious Diseases Research Collaboration (IDRC), Kampala, Uganda
- Makerere University College of Health Sciences, Kampala, Uganda
- Liverpool School of Tropical Medicine, Liverpool, UK
- London School of Hygiene and Tropical Medicine, London, UK
- Radboud Institute for Health Sciences, Netherlands
- Stanford University, Stanford, CA
- Durham University, Durham, UK
- Institute for Health Metrics and Evaluation, Seattle, WA