Asymptomatic Plasmodium infection in a residual malaria transmission area in the Atlantic Forest region : Implications for elimination

Introduction: Elimination of malaria in areas of interrupted transmission warrants careful case assessment to avoid the reintroduction of this disease. Occasional malaria cases are reported among visitors of the Atlantic Forest area of Brazil, while data on residents of this area are scarce. Methods: A sectional study was carried out to examine 324 individuals living in a municipality where autochthonous cases were detected. Results: Asymptomatic Plasmodium infections were detected in 2.8% of the individuals by polymerase chain reaction (PCR), with one case of P. falciparum (0.3%), two cases of P. vivax (0.6%), and six cases of P. malariae (1.9%). The thick blood smears were negative in all individuals. Serological tests performed in 314 subjects were reactive in 11.1%, with 3.5% for P. falciparum and 7.7% for P. vivax. A subsample of 42 reactive individuals for any Plasmodium species showed P. malariae in 30.9% of specimens. Individuals who entered the Atlantic Forest region were 2.7 times more likely to exhibit reactive serology for P. vivax compared with individuals who did not enter this region (p<0.05). Children <15 years had a higher chance of reactive serology for P. falciparum and P. vivax than individuals ≥15 years of age (p<0.05). Individuals living in the Paraiso district had a higher chance of reactive serology for P. vivax compared to other districts (p<0.05). No associations were found between sex, past exposure to malaria, or serological response to antibodies of any Plasmodium species. Conclusions: The implications of these results for the elimination of malaria were discussed.


INTRODUCTION
Malaria endemic areas are restricted to the Amazonian region of Brazil where Anopheles darlingi Root 1926 is the main vector implicated in transmission 1 .An average of 1,296 cases are reported annually outside of this region, most of which (89%) have been imported from elsewhere 2 .While 50% of the malaria cases in the extra-Amazonian region of Brazil were autochthonous in the middle of the 20th century, this proportion has dropped to only 0.05% to date 2 .Rio de Janeiro state, located in Southeastern Brazil, used to be considered a highly endemic area for malaria with the reported presence of An. darlingi 3 .Cases of infection by P. falciparum and P. vivax occurred particularly in the lowlands (Baixada Fluminense region) in the proximity of the capital of the state 4 .The Global Malaria Eradication Campaign supported the implication of control measures against vectors and the discovery of new drugs for infected people, leading to the interruption of transmission by 1968 with An. darlingi now rarely found in this area 5 .In recent years, a new malaria epidemiological scenario is emerging in Rio de Janeiro, with the presence of cases imported from endemic areas of Brazil and outside the country and occasional autochthonous cases including outbreaks 6,7,8,9 .In 2016, the World Health Organization (WHO) launched the "Global Technical Strategy for Malaria 2016-2030", with the goal of "preventing a resurgence of malaria in all countries and areas that are malaria-free" 10 .A recent malaria outbreak caused by P. simium was described in tourists visiting some regions of the dense and protected Atlantic Forest in Rio de Janeiro state 9 .Malaria cases in this area are detected predominantly among visitors but it is unclear whether the local residents are also affected.We therefore carried out a cross-sectional study to verify the status of Plasmodium infections in residents of the Atlantic Forest area of Rio de Janeiro state.The novel epidemiology of malaria in interrupted transmission areas must be carefully assessed, particularly considering worldwide efforts to eliminate the disease 2,10 .

METHODS
A cross-sectional study was carried out during the fall of 2011 in Guapimirim, a municipality located 86 km away from Rio de Janeiro, Brazil.Six localities were eligible for participant recruitment because nearby autochthonous cases were detected, the rural areas Garrafão, Orindi, and Paraíso and the peri-urban areas Barreira, Caneca Fina, and Monte Olivete (Figure 1).The required sample size was estimated using an expected prevalence of 4% for Plasmodium infection as a variable with an acceptable error of 2%.The final sample size was calculated as at least 312 individuals.Residents of both sexes older than 5 years of age who gave their informed consent (for adults) or received consent from the legal guardian (for children under 18 years) were randomly recruited.A semistructure questionnaire on personal data, exposure to malaria, daily habits, knowledge about disease, clinical information, and history of febrile symptoms in the month preceding the study period was administered.Thick and thin blood smears were collected from each participant according to standard protocols of the Brazilian Ministry of Health.Total blood (5 mL) was collected in Vacutainer® tubes (Becton Dickinson New Jersey, USA) containing ethylenediaminetetraacetic acid (EDTA) for DNA extraction for parasitological diagnosis of malaria by polymerase chain reaction (PCR).Ten milliliters of blood were collected without anticoagulant for serological studies.

Parasitological diagnosis
Thick blood smears were stained and 100 fields examined at 1,000× magnification 11 .DNA was extracted for molecular diagnosis from 200 μL of whole blood using a commercial DNA Purification Kit (Illustrablood genomicPrep MiniSpin Kit), according to the manufacturer's instructions (GE Healthcare, Pittsburgh,USA).PCR was performed following the nested PCR protocol by Snounou et al. with minor modifications 12,13 .Agarose gel electrophoresis was performed, gels stained with GelRed TM nucleic acid gel stain, Biotium, Fremont, CA, USA and visualized under UV light.A sample was considered positive if a 120, 144, or 205 bp PCR product (P.vivax, P. malariae, and P. falciparum, respectively) was detected.A sensitivity level of 0.001% for PCR detection of parasitemia levels is appropriate for the diagnosis of subpatent infections 14 .Positive results were re-tested twice to verify the results.

Serological analysis
Serological examinations were performed using enzymelinked immunosorbent assay (ELISA) with erythrocytic antigens for P. falciparum and P. vivax and immunofluorescence assay (IFA) for P. malariae.IgG antibodies were detected by ELISA using a crude blood stage P. falciparum antigen extracted with Zwittergent ® (Calbiochem, Billerica, MA, USA) 15 and MSP1-19 recombinant antigen of P. vivax 16 .Reactions were assessed by measuring the absorbance at 492 nm using Titertek Multiskan MCC/340 (Labsystems Diagnostics Group, Vantaa, Finland).To determine the cut-off for the ELISA using P. falciparum and P. vivax antigens, receiver operating characteristic (ROC) curves were constructed based on the absorbance of positive and negative samples 17 (Figure 2).The reactivity index (RI=absorbance/cut-off) was calculated for all samples, and samples with RI ≥1.1 were considered positive.Samples with values between 0.9 and 1.1 (gray zone) were considered inconclusive.IgG antibodies against P. malariae were detected using the IFA protocol described by Ferreira and Sanchez 16 .As the quantity of P. malariae antigen was low, this serological test was performed only for the PCR-positive samples and all reactive and inconclusive serologies.The experiments were performed at the Laboratory of Seroepidemiology and Immunobiology at the Institute of Tropical Medicine of São Paulo, Brazil.
A malaria case was defined as an individual with any of the typical malaria symptoms (fever, chills, sweating, headache) at the time of the interview and with a positive thick blood smear or PCR result.An asymptomatic Plasmodium carrier was defined as an individual with Plasmodium species detected in the thick blood smear and/or the PCR but without symptoms 30 days before and after sample collection and without using antimalarial drugs.

Ethical considerations
All the procedures followed the ethical standards of the Ethics Committee of Research involving Human Subjects of the Research Institute Evandro Chagas-INI/Fiocruz, Brazil and in accordance with the Helsinki Declaration (Protocol 0229.0.000.009/10approved).

Statistical analysis
Exploratory analyses were performed using contingency tables and chi-square and Fisher tests to verify possible relationships between the dependent and independent variables.This relationship was modeled using a series of simple binomial generalized linear models (GLM).Only variables with p<0.2 in the bivariate analysis were evaluated in the model.Multiple binomial GLMs were used to model the effects of all independent variables on each dependent variable.Interaction effects were tested in each model formulation, but no significant interaction was found.We employed the Firth's Bias-Reduced Logistic Model to account for numerical problems and model convergence in the P. falciparum and P. malariae models.Data were analyzed using the freely available EPI2000 statistical program (Centers for Disease Control, Atlanta, Georgia, USA),

Demographics aspects
We interviewed a total of 324 individuals [186 women (57.4%) and 138 men (42.6%)], with a mean patient age of 32.6± 9.8 years without differences between women and men (p=0.16).Of these, 215 participants (66.4%) lived in rural areas and 109 (33.6%) in peri-urban areas.Most individuals were born in the state of Rio de Janeiro (260/324, 81.8%).The mean educational duration of the participants was 6.9±3.6 years, and 269 people (85.4%) had attended school and could read and write.The average time of residence in the Guapimirim municipality was 18.6±13.1 years, the time of residence in the respective locality was 17.7±12.8years, and that in the current home was 9.8±12 years.A total of 36 individuals (36/324, 11.1%) had changed housing in the last 5 years, 20 of which (6.2%) moved within the municipality.

Previous malaria episodes
Of the 324 participants, 316 (97.5%) never had malaria, 4 (1.2%) did not remember, and 4 (1.2%, 2 women and 2 men) had previously suffered a single episode of malaria.One of these women (age: 69 years) had malaria at the age of 27 years while staying in Mozambique and the other (age: 56 years) had malaria at the age of 6 while staying in the municipality of Mage (state of Rio de Janeiro).One of the previously infected men (age: 70 years) had malaria at the age of 6 years but did not remember the place, while the other man (age: 89 years) had malaria due to P. falciparum infection at the age of 59 years while staying in Rondônia (Amazonian region in Brazil).None of these individuals had symptoms associated with malaria in the year preceding the survey and all of them had negative results in the thick blood smear and PCR tests.

Exposure to malaria
Seventy-eight participants (24.1%) had left the Guapimirim municipality in the 6 months prior to the study, most of them during the last 15 days (41 subjects, 12.7%), and 2 residents went to the Amazon region.Only 10.8% of the participants (35/324) used a mosquito net while sleeping, with 7.7% (25/324) using it always.In terms of visiting the forest region for activities such as collecting fruits, walking, working, or leisure, 75.9% (246/324) of the participants said that they did not regularly enter the forest, 3.1% (10/324) reported going occasionally (mountain climbing or walking), 10.2% (33/324) reported going often, 9.9% (32/324) reported going daily because their homes were in the forest, and 3 participants did not answer this question.Most of the participants (73%) had insufficient knowledge of malaria transmission and prevention.

Laboratory tests
None of the subjects had a positive thick blood smear result.Diagnostic PCR was performed with the samples from 320 (98.8%) individuals.Of these, 9 (2.8%) were positive for Plasmodium infections, 1 for P. falciparum (0.3%), 2 for P. vivax (0.6%), and 6 for P. malariae (1.9%).These 9 individuals did not have symptoms suggestive of malaria and all had negative smears.The individual with a positive PCR result for P. falciparum was a man from the Orindi municipality, and those participants with a positive PCR result for P. vivax were 2 women from the Paradise municipality.The 6 participants with positive PCR results for P. malariae, 3 women and 3 men, were residents of the Orindi, Garrafão, Paradise, and Monte Olivetti locations (Table 1).Although the towns of Paradise and Orindi showed a higher frequency of positive samples, there was no statistically significant association.
Table 4 shows variables associated with IgG reactivity.Individuals who entered the Atlantic Forest region (for hunting, leisure, collecting plants, or other activities) had a 2.7 times increased probability of having a reactive serology for P. vivax compared with individuals who did not enter the forest (p<0.05).On the other hand, children <15 years of age had a higher chance of reactive serology for P. falciparum and P. vivax compared with individuals ≥15 years of age (p<0.05).Individuals living in the Paraiso district had a higher chance of reactive serology for P. vivax and any Plasmodium species (but not specifically for P. falciparum and P. malariae) than people living in other districts (p<0.05).There were no associations between sex, symptoms, or past exposure to malaria and serological response to antibodies of any Plasmodium species.Individuals who reported previous episodes of malaria were not reactive to any species of Plasmodium.

DISCUSSION
The prevalence of autochthonous malaria is low in Southeastern Brazil and in the Southern states of the Atlantic Forest region 18,19,20 .The characteristics of malaria in these areas are different from those observed in the Amazon region 18 .The extra-Amazonian malaria is often described as asymptomatic,  with low parasitemia and predominance of the species P. vivax 20,21,22,23 .The municipality of Guapimirim is very near to the mountain ecosystem of the Atlantic Forest region, and our study was carried out in areas where malaria cases had been reported.Our results showed that the majority of the study population had lived in the municipality for the last 18 years and only 2 participants (0.6%) had a history of travel to an endemic area but without acquiring malaria.Only 1.2% had a single previous malaria episode in the past, but all of them showed negative PCR, thick blood smear, and serology results.
In contrast to our expectations only one of these people had contracted malaria in the Amazon region Rondônia.We observed no human malaria symptomatic cases and all thick and thin blood smears were negative for Plasmodium species.These results are not unexpected because data from the governmental surveillance system have demonstrated that cases  Cohen's kappa index (κ) for associations between molecular and serological results was 0.07 (range: -0.07-0.20,CI: 95%, p>0.05) for P. vivax, -0.07 (range: -0.02-0.01,CI: 95%, p>0.05) for P. falciparum, and 0.542 (range: 0.27-0.82,CI: 95%, p<0.05, "moderate") for P. malariae.The same calculations were performed for all Plasmodium species and identified a kappa index of 0.29 (range: 0.11-0.46,CI: 95%, p<0.05, "fair").
occur throughout the year, with an annual average of 4 or fewer cases across the state particularly during the warmer season 7,23 .Interestingly, malaria outbreaks associated with P. simium occurred in areas close to the municipality of Guapimirim in the summers of 2015 and 2016 9 .In our study, 9 individuals (2.8%) with asymptomatic Plasmodium infection were diagnosed using PCR analysis, 2 of which were positive for P. vivax (0.6%), 1 (0.3%) for P. falciparum, and 6 for P. malariae (1.9%).None of these individuals developed clinical malaria within 30 days of follow-up.Although most of the infected individuals live in the towns of Orindi and Paradise, there were no significant differences between the assessed locations, probably due to the small number of positive samples.It is important to note the presence of asymptomatic infection in individuals with no previous history of malaria, travel to endemic areas, or contact with a previous malaria-infected person.We speculated that these subclinical infections were probably autochthonous cases.Only a few studies on asymptomatic infection by Plasmodium species in the extra-Amazon region have been carried out to date.Curado et al. 24 26 .In our study, we identified 2.8% of participants with asymptomatic infections, 6 of which were infected with P. malaria, the most frequent parasite associated with TTM in the Americas.Thus, further studies should be conducted to identify risk factors in these asymptomatic individuals, to establish diagnostic methods, and to prevent TTM.
In contrast to asymptomatic Plasmodium infections that can only be detected by PCR, the presence of anti-Plasmodium antibodies can reveal recent or past infections 27 .In our study, 11.8% of specimens were reactive to any of the tested antigens, with 3.5% reactive to P. falciparum, 7.7% to P. vivax, and 30.9% of a subsample of 42 individuals reactive to P. malariae.The frequency of reactivity was low in relation to the study of Azevedo (1997, unpublished data) in which 47.8% of samples from the outbreak of 1993 in the Rio Bonito district showed reactivity of IgG antibodies and 17.4% of IgM antibodies (indirect immunofluorescence).These studies were repeated in the area in 1996 with a higher number of subjects and revealed a frequency of 35.4% of IgG antibodies against the asexual blood forms of P. vivax, while no IgM antibodies were found.These findings were corroborated by Mattos et al. 6 Cerutti et al. 20 performed a cross-sectional study of 65 patients and all specimens showed a positive reaction to all the variants of antibodies against the circumsporozoite protein of P. vivax, with positive reactions to P. vivax classic (VK210) in 25.4%, to VK247 in 6.3%, to P. vivax-like in 10%, and to P. malariae in 15.1% specimens.The same authors assessed healthy individuals that had been in contact with malaria patients and observed a high percentage of antibodies to P. vivax (37.7% IgG and 6.2% IgM) and P. malariae (44.6% IgG and 15.8% IgM) and a prevalence of P. falciparum reactivity of 13.5% for IgM and 13% for IgG antibodies against asexual forms of the parasite.With our results, we cannot conclude that serology can predict infection.No previous study has reported a similar trend, and our findings may present a novel epidemiological scenario in this context that needs to be elucidated in future studies.
Enhancing the knowledge on malaria in residents of the extra-Amazonian areas is an important aspect of disease elimination.Although transmission in this area was interrupted 50 years ago 5 , 73% of actual residents have insufficient knowledge on malaria transmission and prevention.Azevedo et al. 6 found a similar percentage in the Nova Friburgo region.It is noteworthy that 79.3% of participants associated the summer with an increase in mosquito density particularly at nightfall, but a large proportion of the population did not use any control measures for avoiding insect bites.Malaria diagnosis is a big challenge in these areas because medical doctors often do not consider the disease when a patient presents with fever 22 .
According to the WHO, residual malaria occurs sporadically in places with interrupted transmission and remaining determinants of transmission 42 .These areas require assessment of favorable locations for Anopheles larvae habitats, food sources, and invasion of Anopheles by Plasmodium.All these conditions are present in the Atlantic Forest region, but on a much smaller scale compared to the Amazon region 36,37,38,39,40 .Albuquerque et al. 43 recently constructed a model for evaluation of territory receptivity in order to strengthen entomological surveillance of Nyssorhynchus mosquitoes for imported malaria cases.The study showed that pluviosity, temperature, geomorphology, and vegetation variables could be used to create a model of surveillance, but a specific model needs to be generated for Kerteszia anophelines responsible for bromeliad malaria in these areas.

FIGURE 1 :FIGURE 2 :
FIGURE 1: Study locations and prevalence of asymptomatic Plasmodium infections in the municipality of Guapimirim, Rio de Janeiro, Brazil.

Miguel
RB et al. -Plasmodium infection in Rio de Janeiro 5/9 05, **Undefined.CI: confidence interval.aOR: adjusted odds ratio, uOR: unadjusted odds ratio. 1 chance of obtaining positive serologies for P. falciparum and P. vivax in individuals under 15 years of age compared to older individuals.

TABLE 1 :
Frequency of Plasmodium infections diagnosed by PCR in the residents of the municipality of Guapimirim, Rio de Janeiro, Brazil, according to the different investigated variables.

Entry to the Atlantic Forest region for hunting, leisure, or collecting plants
Pos: positive; Neg: negative.

TABLE 2 :
Positive serological reactions for IgG anti-Plasmodium species antibodies in the population of the municipality of Guapimirim.
*IgG for P. malariae was assessed only for samples positive in the PCR or serological tests (42 samples).

TABLE 3 :
Relationship between PCR and serology results of individuals in the municipality of Guapimirim, Rio de Janeiro, Brazil.
26monstrated the presence of asymptomatic infections with P. falciparum and mixed malaria caused by P. falciparum and P. vivax in the Atlantic Forest region in São Paulo state.Cerutti et al.20conducted a population-based study in the Espírito Santo state and obtained similar results.No parasites were detected by microscopy in 1,777 blood samples from residents of the area with reported malaria cases, but asymptomatic infections with P. vivax (1.5%), P. malariae (0.9%), and P. falciparum (0.5%) were diagnosed by PCR.De Alencar et al.25in Espírito Santo state found a prevalence of 3.4% of P. malariae and P. vivax infections.Maselli et al26revealed a prevalence of P. falciparum (5.14%) and P. vivax (2.26%) infections in healthy blood donors in São Paulo state.Donors with asymptomatic Plasmodium infection could be reservoirs of transfusion-transmitted malaria (TTM).As these individuals have low parasitic infections with negative thick blood smear results, detection can be difficult using routine laboratory tests Miguel RB et al. -Plasmodium infection in Rio de Janeiro

TABLE 4 :
Reactivity associated with IgG anti-PvMSP119, anti-Pf IgG, and anti-Pm IgG in specimens from residents of the Guapimirim municipality, Rio de Janeiro, Brazil. P.