Located in the eastern portion of the Southeast Region of Brazil, RJ) occupies an area of 43,777.954 km², and is comprised of 92 municipalities (Figure 1). It has the largest population density in Brazil (365.23 inhabitants/km²), with an estimated population, in 2015, of 16,550,024 inhabitants. It is the most urbanized region of the country, with 97% of the population living in cities (IBGE, 2015).

Epidemiological data were obtained from the Secretary of State of Health of RJ (SES/RJ) and from the Sistema de Informação de Agravos de Notificação (SINAN) (Brasil, 2015b), for the time frame from 1980 to 2014. The data were chronologically divided into two periods: Period 1 corresponding to the years from 1980 to 2004; and Period 2 corresponding to the years from 2005 to 2014. This separation of data was related to the epidemiological dynamics of reported cases. In the second period, epidemiological descriptive data (time, place and person) are included in SINAN reporting forms.

Information regarding potential vectors of SF was obtained from the database of the Laboratório de Referência Nacional em Vetores das Riquetsioses (LIRN; Labortory of the National Reference of Rickettsial Vectors), built from samples analyzed and received within the flow of the Rede Nacional de Vigilância de Ambiental para Febre Maculosa e outras Rickettsioses (National Network for Environmental Monitoring for Spotted Fever and other Rickettsial Diseases), of Ministry of Health from 2005 to 2014. The samples were collected in sampling units (specimens originating from the same host or environment, collected on the same date and during the same epidemiological outbreaks) by state and municipal health teams from RJ for environmental monitoring and investigating cases of SF.

In general, the samples were collected by cloth-dragging, visual searches on hosts and inspection of plant litter and abiotic surfaces.

Ticks were identified using dichotomous keys for larvae (Amorim and Serra-Freire, 1999), nymphs (Martins et al., 2010), and adults (Aragão and da Fonseca, 1961; Barros-Battesti et al., 2006), or specific descriptions (Maroun et al., 1999; Walker, 2000). Fleas were identified according to Linardi and Guimarães (2000), and lice according to Ferris (1951) and Price et al. (2004).

Voucher specimens were deposited in Coleção de Artrópodes Vetores Ápteros de Importância em Saúde das Comunidades – CAVAISC/FIOCRUZ, while the remaining specimens were submitted to total DNA extraction by salt-extraction technique (Aljanabi and Martinez, 1997). Species of Rickettsia were identified using PCR. In this process, samples were analyzed as single or pooled specimens ranging from 2 to 10 individuals of the same sampling units, stage of development (larva, nymph, and adult), and sex. For PCR, we used Rickettsia genus-specific primers targeting the gltA gene (CS4 239/CS4 1069, CS2 78/CS2 323) (Labruna et al., 2004) and SFG-specific primers targeting the ompA gene (Rr 190.70p/ Rr 190.602n) (Regnery et al., 1991).

Amplified DNA fragments were visualized using 2% agarose gel stained with ethidium bromide and observed using a gel scanner with ultraviolet light (Sambrook and Russell, 2001).

Previous research (Gehrke, 2010; Moura-Martiniano et al., 2014) used ectoparasites from SF research and surveillance areas in RJ, collected throughout 2005–2009. In these studies, amplicons that corresponded to the expected amplified size were purified and sequenced. The resulting sequences were identified by comparison with the GenBank database using BLAST. These sequences were also submitted to phylogenetic reconstruction. All obtained sequences were deposited in GenBank (Supplementary Table 1).

All information obtained was tabulated according to the type of origin (investigation of cases or regular environment surveillance of SF) and submitted to descriptive statistical analysis; frequency measures, in the case of epidemiological data; and analyses of the percentage of similarity – SIMPER (Clarke, 1993) for biological data. SIMPER was used to highlight species with the highest contribution to abundance among the municipalities of RJ using standardized data of richness and abundance of vectors and the association measure of Bray–Curtis using the statistical package Plymouth Routines In Multivariate Ecological Research – PRIMER-v.5 (Clarke and Gorley, 2006).

To identify statistical patterns of association between parasites (dependent variable) and hosts (independent variable), biological data determined most representative by SIMPER analyses were used for exploratory non-metric multi-dimensional scaling (NMMDS) (Clarke, 1993) using PRIMER-E. Data were standardized with an association measure of Bray–Curtis. In these analyses, the Bubble value of each association is graphically represented on a Cartesian plane by the size (number of individuals) and frequency (number of samples) of each species of ectoparasite collected from each host. For visual comparisons, Bubble groups of each analysis were allocated in the same plane by using ruler measurements and the line grid of the Paint program of Microsoft Office package.

Using the program TerraView (INPA, 2010), and data by area, cartographic analyses were performed using epidemiological thematic maps, with vectors and natural infection with Rickettsial genes to determine distribution, probabilistic patterns and the relationship between these variables using kernel density estimators.

Nine hundred and ninety suspected SF cases were reported during the study period in 73.91% (68/92) of the municipalities within the study area. Of the 990 cases, 11.71% (116/990) were confirmed, including patients residing in 42.30% (39/92) of the municipalities with reported SF cases (Figures 2A,B). Five municipalities (Barra do Pirai, Paraty, Petrópolis, Portiuncula, and Rio de Janeiro) possessed 52.59% (61/116) of the confirmed cases. The municipality of Rio de Janeiro, the State’s capital, had the highest rate, with 18.97% (22/116) of the cases (Figure 2B). In 64.10% (25/39) of the municipalities with confirmed cases, there were 48 deaths (Figure 2C).

The epidemiological dynamics of SF in the study area shows chronologically different profiles (Figure 3). In the first period, between 1980 and 2004, records were mainly quantitative with little additional information. During this period, there were 41 reported cases, with 9 being confirmed and 14 reported deaths.

In the second period, between 2005 and 2014, 949 cases were reported, with 28.11% (107/949) being confirmed and 37.38% (40/107) advancing to death. Confirmed cases were distributed throughout the year, with higher frequencies during the months of July to October. The reported cases of this period also possessed a greater amount of descriptive data. From 2007 to 2014, in 28.95% (22/76) of the cases, the infection occurred in an urban area, 13.15% (10/76) in a peri-urban area and 35.53% (27/76) in a rural area, with the area of infection being unknown in 22.37% (17/76) of the cases.

Households infections occurred in 27.63% (21/76) of the cases, while 22.37% (17/76) occurred during leisure activities and 17.11% (13/76) during labor activity, while 32.89% were not specified.

Males were more frequently infected than females, accounting for 61.84% (47/76) of the infections. In terms of age, the population of reproductive age and the work force, between 20 and 59 years, was the most compromised (57.33%) age group. Even within that period (2007–2014), we found eight cases of SF caused by R. rickettsii (five turned to death, one progressed to healing and two ignored) and one case caused by R. parkeri (progressing to cure).

Distributed among 650 samples, 8,064 specimens of 14 species of wingless arthropods of interest to public and animal health were identified.

Ticks (Acari: Ixodidae) had the greatest abundance and species richness. In order of contribution, the species are: Amblyomma sculptum, Rhipicephalus sanguineus, Rhipicephalus microplus, Dermacentor nitens, Amblyomma aureolatum, Amblyomma dubitatum, Ornithodoros spp., Amblyomma ovale, Haemaphysalis leporispalustris and Amblyomma longirostre (data not shown). The representation of Insecta was dominated by the flea Ctenocephalides felis, with Polygenis spp, Ctenocephalides canis occurring in very low frequency, and the lice Felicola felis and Pediculus humanus occurring in even lower frequency (data not shown).

Of the 37 municipalities with samples sent to LIRN, 30 occurred during the investigation of cases, 10 of which were also part of regular environment surveillance of SF. Seven municipalities had information only for environmental monitoring.

According to SIMPER analysis, the structure of the community of arthropod-carriers of Rickettsia spp. is heterogeneous in the study area, and independent of the origin of the samples (investigation or monitoring). Four species of ticks (A. sculptum, Rh. sanguineus, Rh. Microplu and D. nitens) and one flea (Ct. felis) contributed to 99% of the total arthropod-carrier abundance. However, the contribution of A. sculptum to total relative abundance was dominant in both foci and environmental monitoring areas with 78.67 and 48.86%, respectively.

Potential Rickettsia vectors were found parasitizing eight mammalian species, Bos taurus (cattle), Canis familiaris (dog), Capra hircus (sheep), Equus asinus (ass) Equus caballus (horse), Felis catus (cat), Hidrochoerus hidrochaeris (capybara) and Homo sapiens (human); and two morphospecies Mula and Bos spp. In the environment, mainly immature stages (larvae and nymphs) of A. sculptum and Rh. microplus were found.

Using the five species that contributed 99% of the total abundance, the NMMDS analyses, with stress values in 2D and 3D of 0, found no statistically significant spatial distribution model for vectors in municipalities. With the same amount of stress, NMMDS found that the distribution of ectoparasites to be modeled by their hosts (Figure 4).

Except for A. sculptum, in areas under investigation or surveillance, most species were associated with a single host. Thus, for example, the horse is responsible for maintaining the populations of D. nitens, B. taurus modeled populations of Rh. microplus and the dog was the host most parasitized by Rh. sanguineus (Figure 4).

The presence of the gltA gene was detected in 26 samples of ticks and 24 samples of fleas. Yet, in 43 samples of ticks and 15 of fleas the ompA gene was amplified. Sequences obtained from some samples of RJ ectoparasites exhibited identities of between 97 and 100% with R. rickettsii and R. felis (Table 1 and Supplementary Table 1). In those species in which R. rickettsii genes were found, the ectoparasites had been collected on their specific hosts. A. sculptum was the only species found parasitizing humans with the presence of genes of the causative agent of Brazilian spotted fever (Table 1).

Probabilistic analyses of the distribution of the species A. sculptum, Rh. sanguineus and Ct. felis, with the presence of genes for SFGR and R. rickettsii, had the same spatial model as the likelihood of cases and deaths in the SF study area (Figure 5). The likelihood of cases and deaths by SF was found to be higher in the municipalities of Metropolitan, Middle Paraíba and Northwest Regions of RJ (Figure 5).