Introduction

Rickettsiae are Gram negative obligate intracellular bacteria that include several zoonotic pathogens distributed worldwide [1]. The Rickettsia genus is divided phylogenetically into four distinct groups [2].

Typhus group Rickettsiae (TGR) consist of R. prowazekii and R. typhi, the agents responsible for epidemic and murine typhus, respectively. Rickettsia typhi is usually maintained in rodents and transmitted by the rat flea Xenopsylla cheopis [1]. Invasive rodents, specifically the black rats (Rattus rattus), and the brown rats (Rattus norvegicus), serve as primary reservoirs. Humans are infected through contamination of disrupted skin, respiratory tract, or conjunctivae with infected flea faeces. Symptoms are non-specific and cover a large spectrum from mild to severe illness. Even though TGR pathogens are ubiquitous, their burden is higher in tropical regions [3].

Spotted Fever Group Rickettsiae (SFGR) encompass 20 Rickettsia species (R. rickettsii, R. conorii, R. africae, etc.…) mostly transmitted by ticks (Dermacentor spp., Amblyomma cajennense and Rhipicephalus sanguineus, etc.…) [2]. SFGR multiply in endothelia of almost all organs causing vasculitis. SFGR infections may result into mild to severe and potentially fatal disease.

The ancestral group consists of R. canadensis and R. bellii, and the transitional group is composed of R. akari, R. australis and now R. felis (formerly classified in SFGR). Among these, only the transitional group is pathogenic to humans [2].

After a reporting of eight autochthonous cases of TGR infection on Reunion island between January 2011 and January 2013 and a subsequent case series [4, 5], we conducted a serosurvey using stored frozen human sera dated 2009 to assess the magnitude of exposure to Rickettsiae in the community.

Main text

Methods

Setting and population

La Réunion is a small tropical island (2512 km2), located in the South Western Indian ocean, 700 km east of Madagascar. Its landscape gives rise to contrasted climates with a mountainous centre separating a humid windward east coast from a dry leeward west coast. The lower portion of this latter consists of a savannah where introduced rodents (R. rattus and R. norvegicus) display distinctly higher flea infestation than in the humid portions of the island [6]. Most of the 816,000 inhabitants present in 2009 lived in the coastal area where the cities lie. The structure by age, gender, and microregion (i.e., 4 administrative regional subdivisions) of the general population is presented in Additional file 1: Table S1.

The study population was a subset of the CoPanFlu-RUN cohort, dedicated to 2009 pandemic flu [7]. Data pertaining to age, gender, and place of residence were retrieved for each individual and compared to the general population (Additional file 1: Table S1).

Serology

Two hundred and forty-one sera were tested using an indirect fluorescent antibody (IFA) assay with commercially available R. typhi and R. rickettsii antigens (IF0100G, Rickettsia IFA IgG®, Focus Diagnostics Inc., Cypress, CA, USA) used as TGR and SFGR antigens [8]. A positive specimen with an IgG titre ≥ 1:64 was considered as indicative of a previous infection. Each positive sample was further diluted to 1:128 for a more stringent definition. The sera were checked for Coxiella burnetii [9], another Gram negative obligate bacterium, formerly classified within the Rickettsiaceae family and now assigned to the Coxiellaceae.

Statistical analysis

Statistical analysis was performed using Stata 14.2® (StataCorp, College Station; Texas, USA). The seroprevalence rate was measured as crude (raw or unadjusted) and weighted to account for the disproportions between the CoPanFlu-RUN subset and the general population. Associations between positive specimens for TGR or SGR and age, gender and geographic subdivision were determined using weighted Chi square tests. Survey-adjusted log-binomial regression models were used to estimate prevalence proportion ratios (PPR) and 95% confidence intervals (95% CI). A P value < 0.05 was considered significant.

Results

The raw seroprevalence of TGR infections at a cut-off of 1:64 was of 16.18% (39/241) and the weighted seroprevalence of 12.71% (95% CI 8.84–16.58%). These figures were respectively of 9.13% (22/241) and 8.24% (95% CI 5.04–11.43%) with the more stringent dilution of 1:128.

In comparison, the raw seroprevalence of SFGR infections at a cut-off of 1:64 was of 18.67% (45/241) and the weighted seroprevalence of 17.68% (95% CI 13.25–22.11%). These figures were respectively of 12.86% (31/241) and 10.44% (95% CI 6.89–13.99%) at a dilution of 1:128.

Altogether, these data revealed that a fifth of the population of Reunion island exhibited a positive serology [22.41% (54/241); 18.15%; 95% CI 13.67–22.63%], indicative of a previous infection or exposure with either one of the two Rickettsiae groups.

Both TGR and SFGR infections were less frequent among youths (< 20 years) than among adults, although prevalence did not progress with age through adulthood (Table 1 and Additional file 1: Table S2, Table 2 and Additional file 1: Table S3). Women and men were similarly exposed. The population living in the western dryer microregion was respectively twofold and threefold more likely to have been exposed to TGR and SFGR than people living in the South.

Table 1 Factors independently associated with typhus group Rickettsiae (TGR) seropositivity in multivariate analysis, Reunion island, 2009 (n = 241)
Full size table
Table 2 Factors independently associated with spotted fever group Rickettsiae (SFGR) seropositivity in multivariate analysis, Reunion island, 2009 (n = 241)
Full size table

At the cut-off value of 1:64, 30 out of the 54 positive sera (55.5%) were positive both to TGR and SFGR. Seven positive sera (13.0%) also reacted with C. burnetii (2 with both TGR and SFGR, 3 with SFGR alone, 2 with TGR alone).

Discussion

Rickettsiae are widely distributed in tropical areas and have been previously reported in seabird ticks, rodent ticks and rodent fleas including on Reunion island [10,11,12].

Herein, we provide the first serological evidence of a high magnitude exposure of the population living on Reunion island to pathogenic Rickettsiae.

Given our serosurvey was conducted on sera sampled nearly 2 years before the clinical identification of the first autochthonous cases and that the youngest subjects were relatively spared, we suggest that the TGR and SFGR antibodies are long-lived and rickettsioses have been endemic on Reunion island, as previously hypothesized for the Amazon Basin of Peru [13]. Together, these elements support the idea that Rickettsia infections are unrecognized or misdiagnosed and account for a substantial proportion of fevers of unknown origin, as suggested by a case report of prior murine typhus infection in a traveller returning from Reunion island [14].

Another point to highlight is that seropositive subjects live predominantly on the western dryer part of the island, which is both coherent with the spatial distributions of rat fleas and clinical cases [4, 5]. Furthermore, Rickettsiae from different groups (R. typhi and R. felis) have been identified in rat fleas in this part of the island [12]. Importantly, R. africae has been isolated from Amblyomma loculosum, a seabird tick in remote areas from human dwellings [11], and its main vector Amblyomma variegatum, a cattle tick, has drastically decreased in the western savannah over the last 20 years [15], which makes unlikely the spill over of R. africae from both seabirds or cattle to humans. So, we suggest that R. felis may be partially responsible of SFGR seropositivity on Reunion island. Interestingly, R. felis, a pathogen that share common features of both SFGR and TGR, has recently been identified in Anopheles gambiae mosquitoes and other non-hematophagous arthropods, which makes the understanding of its transmission pathways very complex with several potential vectors [16]. Another possibility is that a slight proportion of SFGR positive serologies may have been acquired while travelling abroad given the western population is also putatively the most mobile on the island. Consistent with this latter hypothesis, SFGR are endemic in the south western Indian ocean region, where both African tick-bite fever, spotted fever and flea-borne spotted fever regularly diagnosed [17, 18].

Finally, our findings warrant a larger scale seroepidemiologic study aimed at better assessing the prevalence of exposure to Rickettsiae and understanding the transmission pathways of these environmental microbial pathogens. Of note, flea indices reported in rodents from the western portion of the island are heterogeneous geographically, and fleas were found only in rats from the dry savannah at a low altitude. Hence, a better powered size study should better discriminate into the habitats and provide a finer resolution of the current epidemiological pattern.

Limitations

This study has potential limitations. First, our findings may be prone to resampling bias given the serosurvey is based on stored frozen sera issued from participants sampled for another purpose. Second, serology cross reactions within the Rickettsiaceae family and with Coxiella burnetii may have precluded the precise identification of the involved pathogens [9]. These things being said, cross reactions with C. burnetii were infrequent (13.0%), so we believe that this limitation is unlikely to change the overall magnitude of Reunion island population exposure to Rickettsiae. Third, it is worth noting that autochthonous clinical cases of spotted fever have never been reported on Reunion island, and within the Rickettsia genus, misclassification may have affected the seroprevalence estimates, given both the likelihood of intra-genus cross reactions and the impossibility to link SFGR antibodies to R. felis exposure in our retrospective study.

Notwithstanding, it would be informative to carry out a more specific serosurvey (such as sero-neutralization) in order to identify the pathogens at the species level.