Resurgence risk for malaria, and the characterization of a recent outbreak in an Amazonian border area between French Guiana and Brazil

After two decades of global decreases in malaria incidence, rates are increasing for the first time since 2016 [1]. In the Americas, the largest increase was recorded in Brazil and Venezuela [1]. Thus, malaria remains a public health challenge in South America. In Brazil, transmission is mainly entrenched in the Amazon Basin, which accounts for 99.5% of Brazil’s malaria burden [2].

French Guiana is a European overseas, malaria-endemic territory. The Oyapock river forms the border between French Guiana and Brazil. This area has engaged in a regional malaria control program [3] which faces challenges concerning illegal populations of gold miners living in remote areas along the borders with Suriname and Brazil [2‐5]. Malaria control interventions are carried out through the free distribution of insecticide-treated nets and access to artemisinin-based therapy on both sides of the border [3]. Along the Oyapock river, P. falciparum has declined over the past several years and P. vivax is now responsible for most malaria cases [6, 7]. Although transmission decline is observed on a global scale, local heterogeneities persist and must be addressed through targeted control initiatives in order to achieve malaria elimination [8, 9]. This is particularly important in border areas, which complicate effective implementation of malaria control interventions [2, 10]. An. darlingi is the predominant malaria vector species in this region [11, 12].

The objective of this article is to describe the 2017 malaria outbreak through an entomo-epidemiological investigation of the Saint-Georges de l’Oyapock (STG) region along the Oyapock river. This investigation is associated with a retrospective study using a spatiotemporal analysis of malaria case data from surveillance health systems of French Guiana and Brazil.

The peak of the 2017 malaria outbreak presented a greater intensity, however a similar pattern compared to previous seasons in both the French Guianese and Brazilian border areas. Several factors may explain this renewed increase in malaria infections. Firstly, there has been a shift from P. falciparum to P. vivax infections in this area over the last decade with greater difficulty (contraindication, availability of G6PD blood test and treatment) in treating with primaquine and thus in avoiding relapses [6, 24]. Secondly, increase in infection rates may be due to environmental causes: the abundance of An. darlingi in Trois-Palétuviers in August–September 2017—which was 7 to 14 times higher than the abundance previously reported from August–September 2013 and 2014 around STG [20] —may result from the magnitude of the rainy and dry seasons in 2017, local geological features and/or the impact of increased deforestation in Amazonia [22, 25]. Thirdly, a new municipal government was inaugurated in Oiapoque in 2017. With that, leadership roles shifted in the Special Indigenous Health District (DSEI), an entity responsible for providing health care to Indigenous peoples in Oiapoque. At the onset of this change in power dynamics, there is often an administrative discontinuity among local governments, which may have consequences for public services [26]. Finally, political and economic regional crises, particularly in Venezuela but also in Brazil, have led to mass human migration into Brazil and is associated with an increase of illicit gold mining [4, 27]. Unexpectedly, our results on incidence and spatial distribution revealed an additional human factor within the specific population of Indigenous South Americans. Study epidemic dynamics show that number of malaria cases began increasing in Indigenous Brazilian villages before increasing on the French Guianese side of the border. These communities were described in a recent study to be particularly at risk for malaria infection, and participated in factors driving this epidemic [5]. Remote Indigenous communities have specific behaviors, precarious living conditions, poor access to health care, higher mobility, as well as commercial, cultural or family links among populations from both sides of the border [5, 13]. This may explain the cross-border epidemic dynamic, and these communities represent a key population for malaria control in the region. Further investigations should be conducted to: (i) clearly identify these hypothetical mechanisms, (ii) confirm that such a scenario may have occurred in previous years, and (iii) show that observed events did not occur independently.

In our study, a little less than 25% of patients (n = 219) were lost in follow-up after chloroquine treatment and didn’t undergo full treatment with primaquine. This result, however, does not change the main results concerning disease incidence. This is a common occurrence with P. vivax infection, specifically in Indigenous villages [24]. Indeed, people often feel better after chloroquine treatment and do not feel the need for additional treatment. Finally, in this study area, neighborhoods, which were more impacted by malaria, tended to be isolated. For the inhabitants of such remote areas, it is expensive and time-consuming to return to a given health center for primaquine treatment [5, 24].

Both contemporary and past studies have shown the significant influence of vector dynamics on malaria transmission [12, 28]. In 2017, the heaviest rainfall hit the area since 2000. This may have impacted the productivity of vector breeding sites in the transitional period between the wet and the dry season, and specifically the abundance of An. darlingi in the dry season [22]. In the specific study region, temperatures are usually not considered to be an accurate predictor of malaria vector presence and density, due to as it oscillates in the ideal range of values for the mosquito development all the time, and present diurnal variations that exceed the seasonal ones [29]. Past studies have demonstrated significant positive correlations between precipitation and An. darlingi abundance [12, 30]. Thus, we expect a significant relationship between precipitation and malaria case incidence. Between 2014 and 2019, during which case number stayed relatively stationary (Fig. 4), a short study performed by the authors showed that annual precipitation explained 91% (r² = 0.91) of variance in the annual number of malaria cases in the STG and Oiapoque municipalities (in the framework of a univariate linear regression). In that sense, the 2017 recrudescence of cases could simply be considered a consequence of climate variability. The relationship between climatic variables and malaria incidence at a finer temporal scale is not trivial. Local ecological conditions [30], control actions, changes in population mobility patterns, etc., are all confounding factors. Eventually, beyond precipitation and temperature, many other climatological variables should be considered (humidity, radiation, etc.) (see for example Adde et al., 2016) [12]. As a consequence, estimating the relative impact of climate variability on malaria epidemiology is highly complicated, and beyond the strict scope of the present work.

Our study of An. darlingi confirm the well-known exo/endophagic behavior of this malaria vector, and the need to prevent host-vector contact inside of households (with mosquito nets) and outside of households (with the use of repellants and loose clothing). Several recent publications refer to insecticide resistance within this species in Brazil [31], and also an increased daytime activity [18]. These factors contribute to the challenge of vector control in the region. Because of the profusion of infected females in August and September—before the human peak in November—, An. darlingi was likely the main Anopheline species involved in the emergence of this outbreak. On the French Guiaese side of the border, most malaria cases occurred in neighborhoods with dense vegetation (Blondin and Trois-Palétuviers) where high densities of An. darlingi were reported both in past and contemporary studies [12, 21]. Blondin and Trois-Palétuviers are also the neighborhoods that are farthest from their proximal health center, 15 min and 45 min by canoe respectively. This isolation may contribute to the foregoing of medical treatment, and may explain the higher rate of relapse in the Indigenous neighborhood of Trois-Palétuviers. In considering the high attack rate in Trois-Palétuviers, our results suggest that the combination of vector density, exchanges and mobility among Indigenous Brazilian villages, isolation, and high relapse rates help to explain the intensity of the 2017 outbreak on the French Guianese side of the border.

Outbreak control measures in French Guiana fostered the use of a specialized malaria team with community health mediators between the months of November and February. This team administered radical treatment to a prodigious fraction of infected individuals. A previous retrospective study in the area did not deploy a similar strategy, and consequently reported less than 5% of infected individuals receiving full treatment [24]. This finding underscores the importance of optimizing response collaborations by integrating community mediators.

The transnational geographic region presented in this study, despite a shared epidemiological setting, possesses two distinct programs for diagnosis and treatment, complicating the implementation of public health programs. Nevertheless, health services on both sides of the border have favorable workforces and technological conditions to positively contribute to successful malaria control. Malaria treatment is free on both sides of the border and “acceptability of interventions” were previously described as “good” in both areas [5]. By countries employ vector control measures, including insecticide-treated bed nets and indoor residual spraying [5]. Brazilian and French Guianese malaria cases appear to be related in space (border area) and time, but also experienced by the same Indigenous populations, notably the Palikur. The Palikur are traditionally a mobile people, commonly traversing the border between French Guiana and the Brazilian villages of Kouméné and Amomi. These two villages are considered the spiritual center for Palikur communities living in Trois-Palétuviers, Philogène or Blondin [32]. Although significant transnational human mobility has been previously documented in this area, more data is necessary in order to better understand local malaria epidemiology [33]. Indigenous health centers are autonomous in Brazil and care is given by and for Indigenous peoples through the National Indian Foundation (FUNAI). Improvements may be needed regarding collaborative efforts between Brazilian health authorities and the FUNAI, in reference to community mobilization, to reach acceptable malaria control metrics in both French Guiana and Brazil. An improved control of malaria in this region requires a global and systematic approach, including partnerships between different disciplines and common commitments among all levels of national institutions.