This study evaluated the prevalence of microscopic and submicroscopic Plasmodium infections in indigenous and non-indigenous communities from Antioquia, Colombia, and its associated factors, to describe the distribution of disease prevalence among heterogeneous populations; this knowledge is necessary to implement proper control strategies for each context [35]. We found that the prevalence of Plasmodium infections was higher in indigenous communities than in non-indigenous communities in both municipalities. Even more, all infections in El Bagre were detected in indigenous communities (11/11), and most of them were asymptomatic and submicroscopic (9/11). On the contrary, most of the infections in indigenous communities in Turbo were symptomatic and microscopic (84.2%).
It is known that malaria transmission in Colombia varies among the endemic regions [35]; in this way, our findings could be explained by differences in malaria profiles in each municipality. Although the general prevalence of malaria in Antioquia has decreased in recent years, the number of cases in El Bagre has been higher than in Turbo (from 190.45 cases/1000 people in 2007 to 21.29 cases/1000 people in 2017 in El Bagre and from 61.53 cases/1000 in 2007 to 0.77cases/1000 people in 2017 in Turbo).
Most of the malaria cases were caused by P. vivax and it is well known that the PQ used for the P. vivax treatment can induce haemolytic crises in individuals with Glucose 6-phosphate dehydrogenase (G6PD) deficiency. The G6PD deficiency is distributed worldwide, however, its frequency varies among regions and ethnic groups. A previous report in malaria-endemic areas of Colombia located on the Pacific coast, found a frequency of G6PD deficiency of 6.56% [36], studies in indigenous populations (Amerindian) are lacking. Considering the high number of P. vivax infections in these populations found in this study, there is a need for further evaluation of the frequency of G6PD deficiency in malaria-endemic areas in view that the primaquine treatment (14 days) is required for the radical cure.
As previously reported, malaria immunity is determined by previous Plasmodium exposure, where an anti-disease immunity is first achieved, resulting in a reduction of severe malaria and mortality. Then, an anti-parasitic immunity is slowly acquired and confers protection against high parasitic densities, which in turn protect against the severe disease [37]; this could explain the highest prevalence of submicroscopic infections in El Bagre, where 50.3% of individuals had had more than one malaria episode over life compared to 35.8% in Turbo (Supplementary Table 1). Nevertheless, we could not find an association of this variable with the Plasmodium infections using a GEE analysis. However, a previous study in Nariño- Colombia showed that having suffered from more than one malaria episode was associated with an increased risk of having asymptomatic infections (aOR 2.4, 95% CI 1.1–5.4) [22]. These differences could be explained because our model included not only asymptomatic but also symptomatic infections.
Household factors are also associated with malaria risk [38]. We found that having no access to electricity was associated with an increase in the malaria rate. These findings are in agreement with previous studies that reported that the poorest households had a 29% greater risk of microscopic parasitemia compared to the poorest houses (aRR 1,29; 95% CI 1,07–1,55) [39]. Additionally, lack of household electricity increased the childhood mortality in Rwanda, including malaria mortality (aOR: 1.4, 95% CI: 1.0–1.8) [40]. The above is important because housing quality can affect malaria risk through its effect on house entry of the malaria vector [39].
Taken together, the individual and housing characteristics could help to understand why the indigenous population has a higher prevalence than its counterpart, the non-indigenous population does. Ethnicity is an important determinant of health conditions, influencing the morbidity and mortality rates in different ethnic groups and interfering with access to health services for some populations [41]. In Colombia, the exclusion of indigenous people is reflected in poverty rates, lack of land and employment, school desertion, unsatisfied basic needs, a higher prevalence of transmissible diseases and limited access to health services compared to the general Colombian population [42]. Regarding this last point, we noticed that the indigenous villages were farther from the health services (1 to 2 hours by motorcycle) compared to the non-indigenous villages, and the road conditions were far worse. Furthermore, the indigenous population frequently lives close to rainforests or wetlands where they have more vector exposure, resulting in an increased risk of getting sick with vector-borne diseases such as malaria [8].
It is possible to suggest that the diversity of epidemiologic characteristics of malarial infection among the Colombian subpopulations account for an ideal environment for parasite evolution. In this environment, the parasite can interact with susceptible populations from different ethnicities and under different public health interventions [35]. We suggest that prevention efforts should be population-specific and vary according to the individual, housing, and environmental characteristics. Given the heterogeneity of the prevalence of malaria in Colombia, it has become necessary to adjust malaria control activities according to each population and context.
This study has some limitations. First, due to cross-sectional design, the association with malaria status should be interpreted with caution, as they do not imply causality. Second, it was no possible to analyze the risk factors for asymptomatic infections exclusively due to the low number of this type of infection. This last could, in turn, affect the accuracy of confidence intervals for some of the factors analyzed due to the sample size. Third, as we mentioned before, the villages in this study were selected based on the historical records of malaria cases, the distance to the urban area, and the accessibility for field staff, and we did not make a random selection of the villages included in the study. The results of this study cannot be extrapolated to the general population; nevertheless, they are useful to exhibit the problems around the asymptomatic infections in the indigenous and non-indigenous people. Fourth, considering that nowadays there are ultra-sensitive molecular tests for the detection of low-density infections, the prevalence in this study could be underestimated due to the limit of detection of the nPCR used, nevertheless, the nPCR used in this study was able to detect 1.6 times more infections than microscopy showing the presence of a significant Plasmodium reservoir in the region. At last, it is possible that other variables which we did not consider in the GEE model could explain the associated factors to Plasmodium infections. Future studies are required to improve the knowledge that we have about the risk factors of the Plasmodium infections in indigenous communities. Despite these limitations, these results are useful to understand malaria transmission in studied places and to suggest prevention efforts according to the individual, housing, and environmental characteristics.