Field Performance of Mass PCR Screening and Targeted Treatment in an Amazonian Low Malaria Transmission Setting

The three main obstacles to the elimination of malaria in French Guiana are asymptomatic carriers of Plasmodium vivax, relapses and, to a lesser extent, Plasmodium falciparum. This study aims to assess the impact of PCR-based mass screening and treatment (MSAT) interventions in this malaria-endemic area. Two MSAT interventions were conducted twelve months apart in inhabitants of Saint Georges de l’Oyapock village, which has the highest malaria burden in French Guiana. Symptomatic malaria incidence was also passively monitored through the local health center from 12 months before the rst intervention until the end of the second intervention. At the time of the rst intervention, malaria prevalence was 6.7% [CI 95 5.4-7.9%], including 90% of Plasmodium vivax cases and 10% Plasmodium falciparum (n=1,501 participants). Twelve months later, it had decreased by 53.7% to a value of 2.5% [CI 95 2.0-3.9%] (p<0.05; n=1,271 inhabitants), of which 83% and 17% of cases showed Plasmodium vivax and Plasmodium falciparum carriage, respectively. Similarly, the passive malaria detection carried out by the health center during the 12-month surveillance period that followed the rst MSAT noted a decrease in symptomatic Plasmodium spp.. This study suggests that the implementation of mass PCR testing and the subsequent malaria treatment of positive cases could reduce the prevalence of both symptomatic and asymptomatic malaria infections in the Amazonian context.


Introduction
Global efforts to scale-up malaria control interventions are essential in the ght against this disease.
Despite the deployment of means such as the test-treat-track strategy, Long-Lasting Insecticide Nets (LLIN) and Indoor Residual Spraying, malaria incidence around the world has increased since 2016, particularly in Amazonia 1 . Although climatic, economic and political factors could partially explain this increase in transmission, a signi cant role is played by asymptomatic parasite carriers who generally fall outside the scope of control interventions.
French Guiana is a malaria-endemic French overseas territory in Amazonia, bordering Brazil. This area is committed to a regional program for malaria control 2 . Over the last ten years, there has been a steady decline in Plasmodium falciparum (P. falciparum) transmission. However, the region now faces an increase in the proportion of Plasmodium vivax (P. vivax) infections 3 . P. vivax biology could explain this situation, namely through the occurrence of dormant liver stages (hypnozoites) in humans and the consequent relapses and early circulation of gametocytes in the blood during the infection. A radical primaquine treatment is required to clear hypnozoites, but this drug is associated with a risk of haemolysis in individuals with glucose-6 phosphate-dehydrogenase (G6PD) de ciency. The administration of primaquine to such patients could lead to severe side effects. G6PD testing in French Guiana is therefore crucial but also di cult to obtain 4 , meaning that primaquine treatment coverage is globally insu cient in French Guiana, and particularly in remote areas where malaria is endemic.
The use of mass drug administration (MDA) to eliminate malaria has been widely evaluated over recent years 5,6 . Although these campaigns generally reduce the initial risk of malaria carriage, few studies describe a sustained impact beyond six months post-treatment. MDAs also show more success in reducing the transmission of P. falciparum than in decreasing P. vivax infection 7 . Mass-Screen-and-Treat (MSAT) approaches have also been evaluated, most of which were associated with rapid diagnostic testing (RDT) due to its ease of use in the eld 8 . MSAT had a limited impact in high transmission settings, but this may be explained by the low capacities of RDTs to detect asymptomatic carriers 9 .
For this study, we hypothesized that MSAT could reduce P. vivax transmission and thus the prevalence of both symptomatic and asymptomatic parasitaemia in our low transmission setting in South America. The MSAT study was based on Polymerase Chain Reaction (PCR) testing and was implemented in two phases, twelve months apart.

Materials And Methods
This pre-post study, named "Palustop", occurred in the Amazonian context of the Guiana Shield and involved two cross-sectional interventions occurring one year apart and a passive monitoring of all malaria cases in the village population by the local health centre. The rst campaign was implemented from October to December 2017, and the second took place one year later from October to December 2018. The methodology, prevalence and spatial distribution of malaria cases characterized during this rst MSAT campaign have already been reported 10 .

Study Area
The Saint-Georges de l'Oyapock (STG) municipality is located beside the Oyapock River, the natural border with the Brazilian State of Amapá (Fig. 1). The neighbourhoods of STG were selected for this study due to their history of high incidence of malaria cases, as recorded by the regional surveillance system in 2015 and 2016.
In this part of Amazonia, the equatorial climate involves four annual seasons: a long rainy season from April to June, a dry season from July to December, a short rainy season from January to February and a short dry season in March. In this village, Anopheles darlingi is the predominant Anopheles species 11 .
However, its density is heterogeneous and varies according to the proximity to dense forest 12 . This vector density increases following every dry season and is followed by an increase in malaria cases a few weeks later 11,13 . Two peaks in malaria cases are generally observed in this area, the rst occurring in June and the second in November 14 .
Population STG village has ~ 4,033 inhabitants (STG health centre data 2017) of autochthonous origin such as Amerindians (mainly Palikur), French Guianese Creoles but also migrants originating mostly from neighbouring Brazil. The single health centre located in the area is the frontline provider of malaria diagnosis and treatment, and distributes LLINs to all pregnant women and positive malaria cases.

Sample size
The entire population size of the targeted neighbourhoods was ~ 2,727 inhabitants (67.7% of the total population), all ages and nationalities included. Based on previous data, and assuming an expected PCRprevalence of 4%, an error-margin and an alpha risk of 0.05, the required sample size was estimated at 1,200 15 . Given the mobility and expected participation of the population in this area, we targeted a sample size of 1,500 participants for the rst intervention, expecting that 1,200 of these individuals would be monitored one year later.

Outcomes
The primary outcome was a reduction in the Plasmodium spp. prevalence values determined by PCR in the second MSAT intervention compared to the rst. The secondary outcomes were the incidence of clinical malaria episodes in participants and non-participants, an increase in the time before the rst recurrence occurred, and the observation of severe adverse events over 30 days after initiation of the antimalarial drug treatment.

Community engagement
Community awareness interventions were conducted during the MSAT intervention to secure the commitment and participation of a majority of the inhabitants. Community engagement started in meetings with the community leaders such as the Amerindian community and the village mayor, as well as healthcare workers and the members of local associations. Stakeholders were sensitized to the study objectives, interventions and expected role of the community. Community engagement meetings were held in neighbourhoods in local languages with trained cultural mediators 16 . Health participants were informed about malaria and the rational of mass testing to limit and prevent malaria infections. Positive malaria cases were treated and followed at home by a mobile team composed of an infectious diseases specialist, a nurse and a cultural mediator who informed patients on how to take their treatment sought to improve medication adherence and reported any side effects. Additional mobilization strategies included media campaigns and WhatsApp messages. At the end of the study, each neighbourhood received feedback about the results.

Biological analysis
Rapid Diagnostic Testing (RDT) for malaria was performed, and any participant with a positive result was immediately treated. The test used was the SD BIOLINE® Malaria Ag Pf/Pan test (pfHRP2/pLDH-based, Standard Diagnostics), commonly used in French Guiana. Malaria diagnosis was performed using a realtime PCR derived from Shokoples et al. and accredited according to the medical biology norm ISO NF EN 15189 17 . Detection and identi cation of the four major Plasmodium species (P. falciparum, P. vivax, P. ovale and P. malariae) were performed using a Taqman probe strategy with a sensitivity of 1 parasite/µl of blood. Blood counts including reticulocyte were performed, and glucose-6-phosphate dehydrogenase (G6PD) activity testing was also carried out.

Interviews and medical care
During each MSAT phase, participants were interviewed by trained cultural community health workers in the language of participants (Creole, Palikur, Portuguese, French) and using a questionnaire 10 . The homes of participants were georeferenced. The temperature of each patient was measured. All participants had a medical examination. If RDT or PCR results were positive, participants were treated according to the regional treatment guidelines 2 . P. falciparum infections were treated by artemetherlumefantrine and P. vivax by chloroquine for 3 days followed by 14 days of primaquine (0.5mg/kg) in the absence of medical contraindications (pregnancy, G6PD de ciency or breastfeeding an infant below the age of 6 months) 18 .

Passive follow-up of cohort
The health centre of STG is the only place where malaria diagnosis and treatment could be carried out in this municipality. Data from the health centre provided exhaustive records for medical consultations in this population. P. vivax relapse was de ned as a patient with a medical history of malaria diagnosed within a period of 7 to 90 days of the last malaria diagnosis. A P. vivax malaria case was considered to be new if it was diagnosed more than 90 days after a previous malaria infection, as de ned in a previous local epidemiological study 19 . This interval was considered adequate to distinguish between recurrences or relapse and a new infection.

Data analysis
Symptomatic cases were de ned as individuals presenting with a positive RDT or PCR result and fever (≥ 38°C) or a history of fever in the 48 hours preceding or including the diagnosis test. Conversely, Plasmodium spp. asymptomatic cases were de ned as Plasmodium-PCR positive individuals presenting without fever or any history of fever in the last 48 hours.
Survival analysis was used to analyse the cumulative incidence of Plasmodium spp. infection over 12 months after the rst MSAT intervention. The difference between the two survival curves was assessed at month 12 through a Kaplan-Meier estimate using log-rank test.

Enumeration of study participants and MSAT coverage
More than half of the targeted inhabitants (1,566/2,727, 57.4%) were included in the study and participated in MSAT1. PCR screening involved 1,501 of these participants (Fig. 2).
Before the MSAT intervention, around 33.7% of the population of STG had a medical history of malaria, mainly involving P. vivax (Table 1). falciparum with 88.8% of asymptomatic carriage) during the 2nd campaign (p < 0.005) (Fig. 2). All participants with positive PCR tests received treatment during the two interventions if there was no contraindication. Any severe side-effect was reported. Of 90 P. vivax cases, ve concerned pregnant women and seven concerned participants with a G6PD de ciency. These patients did not receive primaquine treatment (n = 12, 13.3%).
Participants carrying P. vivax at the 1st MSAT were more likely to have a P. vivax recurrence before and after the MSAT campaign (p < 0.005, Table 2).  14 . To evaluate its impact on the results of this study, we compared malaria incidence between the two MSAT interventions through the passive monitoring of study participants and non-participants (supplementary Fig. 1). A signi cant decrease in malaria incidence was observed in the study participant group (95 to 43/1,566 person-years, p < 0.005) between the two MSAT campaigns but not in the non-participant group (38 to 59/2,467 personyears p = 0.24) (Fig. 3).
The passive monitoring of symptomatic cases in the municipality of STG recorded 235 malaria cases (212 P. vivax, 8 P. falciparum and 15 Plasmodium spp. cases) (Supplement le Fig. 1). Between October 2017 and December 2018, the variations of malaria incidences over time were similar among MSAT participants and non-MSAT participants living in the study area and even outside these neighbourhoods (Supplement le Fig. 1). However, variation was higher in the targeted neighbourhoods, as expected.
Analysis of Kaplan-Meier curves revealed also a signi cantly greater decrease in P. vivax infections in MSAT participants than in non-participants (log-rank p<0.05) after the campaign (Figure 4). During the same period, there was no signi cant difference in the Kaplan-Meier survival curve of P. vivax relapse between participants and non-participants of the rst MSAT intervention (log-rank p=0.99) (supplementary g 2).
Only four cases of P. falciparum infection were reported during the follow-up period, namely three MSAT participants and one non-participant. G6PD de ciency and contraindication All P. vivax or P. falciparum carriers received treatment and were monitored. Following the WHO de nition, 9.2% of the participants presented a severe G6PD de ciency (n=13/136 below 30% of the normal G6PD activity 20 , Table 3) and were treated with chloroquine alone. No severe side effects were reported.
No signi cant differences in the number of malaria infections were observed between MSAT participants with G6PD de ciency, meaning without primaquine treatment, and participants without G6PD de ciency in the number of malaria infection recorded between the two MSAT interventions (Table 3).

Discussion
This study provided evidence of a sustained decrease in the prevalence of Plasmodium spp. carriage over at least one year following an MSAT campaign. The MSAT also noticeably reduced the incidence of symptomatic P. vivax infections in participants for more than 12 months. Relapses, and to a lesser degree reinfections, are the most likely source of P. vivax infections, and are a source of ongoing transmission. In French Guiana, P. vivax infections generally relapse after a short interval (up to 90 days, and generally around 28 days) 19 . Therefore, the MSAT campaign including a radical cure of chloroquine and 14 days of primaquine will block transmission by clearing P. vivax gametocytes and also hypnozoites. Nearly half of the P. vivax cases in the studied area were previously reported to be caused by latent hypnozoites in patients who had not bene ted from primaquine as a radical cure 4 . The obstacles to primaquine administration were mainly logistical (samples for G6PD testing were shipped by canoe and drugs delivery followed a complex pharmaceutical approval process) 4,10 . In addition, patients feel better after chloroquine treatment, and do not usually return to the health centre for G6PD testing and primaquine treatment. For these reasons, this MSAT campaign with dedicated teams including cultural health mediators increased primaquine delivery to 100% of P. vivax carriers who had no contraindication for its use. This approach was rapid and suitable for local populations, including a personalized follow-up. Despite this, only 60.7% of the primaquine-treated cases reported adherence to the complete 14-day course. This is a long period of treatment and follow-up did not include a daily visit to supervise primaquine intake. This poor adherence could explain why relapses were observed in MSAT participants.
Single-dose tafenoquine, which was successfully evaluated in neighbouring Brazil, could be a solution to increase medication adherence 21 . However, tafenoquine involves stronger side-effects in case of G6PD de ciency than for primaquine, and this treatment is not yet available in French Guiana.
Several Mass Drug Administration (MDAs) including mass primaquine administrations have been used to eliminate P. vivax throughout the world 22 . When MDA achieved high coverage and was associated with vector control measures, it was described as an effective means to decrease incidence 22 . However, data are scarce about its e ciency on asymptomatic carriage and therefore on the reservoir 9,23 . MDA of primaquine showed a good tolerability, despite high levels of G6PD de ciency in some regions of the world 22 . In our study, participants with severe G6PD de ciency and pregnant women were only prescribed chloroquine treatment in cases of P. vivax carriage. However, these populations did not reveal signi cantly more relapses or carriage in our study. The di culty to detect such a difference could be explained by the lack of statistical power, given the low number of patients with G6PD de ciency in the study samples. This study not only allowed a reduction of Plasmodium spp. carriage, but also enabled us to determine the G6PD status of a large part of this population living in a high malaria-risk area. These values have been recorded carefully in the health centre. This will enable participants to bene t from a faster primaquine treatment administration in the future.  25 . The treatment in the study by Gar eld et al. was also shorter than that prescribed in our study, with a three-day course of chloroquine and a veday course of primaquine (15mg/day) 24 . However, P. vivax carriers in our study had a higher risk for P. vivax infection and a shorter median time before the next infection compared to participants without P. vivax carriage. This could be due to a higher global risk for malaria when living in speci c areas such as a remote neighbourhood, or socio-demographic parameters such as high malaria-risk activities (hunting, shing or farming in the forest), as previously described 10 . It is also important to note that although primaquine coverage was 100% at the rst MSAT intervention, medication adherence to 14 days of primaquine was low (around 60%); this gure should be taken with caution due to a possible memory bias (patients were asked about primaquine adherence one year after treatment). Similarly, data regarding behaviours in cases of fever during the 2 nd MSAT identi ed a lack of medical consultation for malaria testing (56% of patients with fever presented for malaria testing), and this despite health education provided by dedicated teams during the 1 st MSAT. In routine care, primaquine coverage was also low In our transborder context, this kind of intervention must be repeated and coordinated with our Brazilian neighbour in order to be effective over the long term, and avoid new infections in this territory. It is also important to target mobile populations. In this respect, the ongoing MALAKIT project seeks to provide a kit for auto-testing and auto-treatment, and aims to control malaria in mobile gold miner populations 26 . Efforts are also being made to develop and maintain a cross-border observatory on vector-borne disease (notably by the French-Brazilian Joint International Laboratory "Sentinela") and a political commitment to eliminating this disease, with regular meetings and the implementation of speci c measures in the eld 27 .
The efforts made to engage the empowerment and participation of the communities at all levels led the local population to accept and willingly participate in the MSAT campaigns. Communication via cultural community mediators who know local culture, languages and community mobilization made it possible to build con dence about MSAT and increased adherence 28 . Door-to-door distribution of medication and the follow-up of carriers were effective in targeting all positive cases and improved the care of malaria patients. It was also easier for cultural mediators to explain directly to patients that the drugs had to be taken regularly, even by those who were asymptomatic and did not feel sick, in order to protect their household and community and eventually defeat malaria transmission. Being close to the inhabitants was also a logical approach given the logistical di culties and the high cost of transport from houses to the health centre in this area. However, primaquine treatment adherence remains particularly challenging.
Although MSAT implementation was a success, it has a number of limitations. MSAT activities were conducted during the day, making it di cult to reach adult men who were at work, hunting or shing. In addition, these communities move around a vast border territory for family purposes (ex: Amerindians moving between Brazil and French Guiana) or for subsistence. Finally, during the rst intervention, a major seasonal malaria epidemic was reported. This outbreak may have resulted in an overestimation of MSAT e ciency. However, results of passive monitoring of non-participants of the study were a good indication of the impact on the number on symptomatic cases of MSAT strategy.

Conclusion
In a border and remote area context, MSAT intervention was feasible, safe, and effective for not only Plasmodium spp. carriage prevalence but also P. vivax infections, with a follow-up of more than 12 months. The elimination of malaria caused by P. falciparum or P. vivax could be accelerated by interventions in the general population, targeting malaria transmission with sensitive diagnostic tools and adapted treatments.