Rapid Viral Expansion Beyond the Amazon Basin: Increased Epidemic Activity of Oropouche Virus Across the Americas

Introduction

Oropouche virus (OROV; Orthobunyavirus oropoucheense) is an arthropod-borne virus classified within the order Bunyavirales, family Peribunyaviridae and genus Orthobunyavirus (1). It is present as a negative-sense, single-stranded RNA genome that is divided into three segments based on size: S (small), M (medium) and L (large). These segments encode four structural proteins: the nucleocapsid, two external glycoproteins, and the RNA polymerase (2). OROV was first identified in 1955 in Oropouche, a village in Trinidad and Tobago (3). Since then, the virus has been responsible for numerous outbreaks, mostly within the Amazon Basin, where it is found among forest animals such as non-human primates, sloths, and birds (4). The midge Culicoides paraensis is the primary vector for human transmission. Infection with OROV causes Oropouche fever, which typically presents as fever, headache, muscle and joint pain (5). The majority of human infections present as mild to moderate disease and resolve within a week, although rare cases can lead to complications like aseptic meningoencephalitis (6).

Historically, OROV outbreaks were largely restricted to the Amazon region, with about 30 outbreaks reported in Latin America until recent years (7). However, new epidemiological data show a marked increase in Oropouche fever cases in Brazil, Cuba (N=74), Bolivia (N=356), Colombia (N=74), and Peru (N=290), with travel-related cases reported in Italy, Spain, and Germany, all originating from Cuba, highlighting that it’s becoming recognized outside the region (8, 9). Notably, Brazil alone reported a total number of more than 7,000 cases this year (up to August 2024) compared to 836 in 2023, indicating a significant rise in transmission (10-12). In 2024, Brazil also reported the first three fatal cases associated with OROV infection, raising concerns among health authorities (13). OROV has also recently appeared in historically non-endemic Brazilian regions outside the Amazon, including in the far south and, importantly, in some urban centers on the east coast (13). As is the case for other arboviruses (14), recent changes in disease ecology, such as deforestation, urbanization, human mobility and climate change, are some of the possible drivers of its recent emergence in the Amazon region (15). Local reservoir habitats belonging to non-human mammals and vectors can be disrupted, pushing them into closer contact with each other and with urban and peri-urban areas where humans can be infected. In addition, human mobility favors long distance viral movement. Additionally, ongoing changes in OROV genetic diversity may also result in changes in virulence and transmission potential (16).

Scientific and surveillance data for OROV are currently limited, with fewer than 110 peer-reviewed publications compared to thousands for other arboviruses such as Zika or Dengue (17). Reports of recent increases in OROV epidemic activity underscore the urgent need for more data and research. Some of the epidemic activity over the past decade has provided valuable insights into OROV epidemiological characteristics, as well as its potential for public health impact. To help fill current knowledge gaps in the face of recent epidemic activity in the south of Brazil we, in collaboration with several Central Public Health Laboratories, generated 133 viral genome sequences including all three segments (L, M, and S). Herein, we present genomic analysis that offers insights into OROV’s recent movement from northern to southern Brazil and its emergence in regions, within and outside the country, classically not associated with epidemic activity.

Methods

Results

Between late 2022 and early 2024, the Brazilian states of Acre, Amazonas, Rondônia, and Roraima, located in the western Amazon region, reported a sharp increase in the incidence of OROV human cases (11). Concurrently, there was a substantial increase in the number of real-time RT-PCR tests conducted across the country, reflecting heightened screening efforts (Figure 1a). In 2020, initial screening efforts focused primarily on the northern region of Brazil, with the Roraima and Pará states accounting for the majority of the 238 tests conducted (31). At the same time, a few positive cases were detected in Amapá, Amazonas, Pará, Piauí, and Rondônia, signaling the early spread of OROV. By 2021, the number of tests surged to 1,466, with significant increases in states outside the Amazon basin such as Minas Gerais and Ceará. Positive cases were reported predominantly in Amapá, Pará, and Piauí (13). The trend of increased testing continued into 2022, with 588 tests conducted, marking notable expansions in Midwestern and Northeastern Brazilian states. In 2023, the screening efforts intensified further, with 5,280 tests conducted nationwide, with large testing numbers in Bahia, Goiás, Rio de Janeiro, and Tocantins, reflecting widespread concern. Positive cases were detected in multiple states, including Acre, Maranhão, Mato Grosso, Pará, Piauí, Rio Grande do Norte, Rondônia, Roraima, and Tocantins (Figure 1a) (13). By early 2024, screening efforts reached an unprecedented level, with 54,428 tests conducted across numerous states (Figure 1a). As shown in the inset panel, the positive rate nearly doubled, increasing from 0.059 in 2020 to 0.1 in 2024. The states of Espírito Santo, Minas Gerais, Bahia, and Goiás reported the highest testing numbers outside the Amazon basin (13). Within this period, cumulative positive cases were highest within Rondônia (n=1,747), Bahia (n=837), Espírito Santo (n=416), and Roraima (n=244) (Figure 1b).

• Download figure
• Open in new tab

Figure 1. Distribution and Epidemiological Insights of OROV Clinical Cases Detected Beyond the Amazon Basin.

a) Weekly notified OROV PCR tests and positive cases normalized per 100K individuals per region from 2020 to 2024. The inset panel shows the positive rate (ratio of positive cases to PCR tests) over the same period, with the y-axis scaled from 0.05 to 0.1 to highlight the variations; b) Map of Brazil showing the number of new OROV sequences by state. The color and size of the circles indicates the number of new genomes generated in this study; c) Number of new OROV genomes per state obtained in this study (Santa Catarina n=3; Acre n=6, Mato Grosso n=9; Bahia n=34; Minas Gerais n=81); d) Number of genomes generated in this study compared to the number of Brazilian OROV sequences available on the GenBank up to 31st July, 2024. Bars are months, months with 0 sampling are not shown; e) Cumulative OROV cases in absolute (full line) and by gender (male dotted, female dashed) of samples from 2024 for which gender metadata was available; the inner panel shows the gender ratio (M/F) per date; f-g) Observed (bars) and theoretical (lines) age-distributions for males (f) and females (g) of samples from 2024 for which gender metadata was available. Fitted theoretical distributions were Weibul (male: shape 2.11, scale 44.39, mean 39.15; female: shape 2.07, scale 44.22, mean 39.16).

Five Brazilian states were represented in the 133 newly generated genome sequences: Acre (N=6, northwest), Bahia (N=34, northeast), Mato Grosso (N=9, midwest), Minas Gerais (N=81, southeast) and Santa Catarina (N=3, south) (Figure 1c,a). This sampling covered a wide spatial range outside the Amazon basin (Figure 1a), representing the most time-intensive sampling period to date in Brazil (Figure 1d). The average cycle threshold (CT) for the three genes was 25, ranging from 8 to 36 (Table S1). The gender ratio associated with the genome samples was biased towards females in February-March 2024, but converged to 1 into late local autumn (Figure 1e). Cumulatively, 53% (n=70) were female and 47% (n=63) were male (Table S1), with genders showing a similar age profile (Figure 1f-g). Ages ranged from 1 to 89 years, with a median age of ∼39 years.

Genome sequences were obtained from all five Brazilian macro-regions (North, South, Northeast, South East and North East), revealing north-to-south viral movement across the country (Figure 1). The most frequent symptoms observed among the patients were fever, myalgia, and headache. These symptoms were consistently reported across multiple cases, with some patients also experiencing arthralgia. In addition to the common symptomatic presentations, our study identified three fatal cases associated with OROV infection. Notably, these fatalities occurred in young adult patients with no reported comorbidities. In all cases, the clinical course resembled severe dengue, with shock, bleeding, and extensive coagulopathy. Without the extensive lab evaluations due to the ongoing OROV outbreak in the region, these deaths would likely have been misclassified as Dengue Fever rather than OROV. The detection of these fatal cases suggests that OROV may have a broader clinical impact than previously understood, warranting further investigation into the factors contributing to severe disease outcomes. Point mutations were identified in all three fatal cases. In the M segment, amino acid changes included I to V at position 13, M to I at position 642, A to T at position 752, and R to K at position 1342. In the L segment, mutations were found at amino acid positions 857 (T to A), 1634 (K to E), and 2206 (N to D). Further in-depth analyses are required to determine whether these mutations play any role in the pathogenesis and severity of OROV infections. There was seemingly no evidence that symptoms varied with age or gender (Figure S1).

The sequencing procedures yielded an average coverage of 97.7% for segment S, 98.5% for segment M, and 98.32% for segment L (Table S1). Using a Generalized Additive Model similar to the one used for other arboviruses (30), we generated response curves for resulting sequence coverage dependent on sample CT. As expected, both observed (Table S1) and estimated sequence coverage were negatively correlated with CT value (Figure 2a). Segment M overperformed and segment S underperformed in sequence coverage independently of CT value (albeit only marginally). Segment M was the least sensitive to CT value, showing consistent sequence coverage across sample CT. The states of Bahia and Minas Gerais (∼26 and ∼61% of samples, respectively) represented extremes of model output, with consistently lower sequence coverage for Bahia and higher coverage for Minas Gerais, independently of sample CT and across all genome segments (Figure S2). Prolonged intervals between symptom onset and sample collection were associated with increased CT values (Figure 2b), suggesting a decline in viral RNA quantity and consequently sample quality over time, underscoring the critical importance of minimizing the time between symptom onset and sample collection to ensure the acquisition of high-quality sequencing data. Additionally, considering the segmented nature of the OROV genome, we investigated the potential presence of reassortment and recombination among the three different genome segments. We identified 21 reassortment events: 17 separating the S and M segments, 7 separating the S and L segments, and 11 separating the M and L segments.

• Download figure
• Open in new tab

Figure 2. Sequencing Coverage and Sample CT Value Analysis: Model Estimates.

a) Summary of estimated sequence coverage dependent on sample CT using a Generalized Additive Model. Shown are the mean (points) and 95% CI (ranges) for model estimates given selected values of CT (10, 20, 30, 40, in color) for all segments separately and together; b) Sample CT values (mean as points, standard deviation as ranges) plotted against the sampling delay (days between symptom onset and sample collection).

To reveal the recent evolution of the three OROV genomic segments, we estimated three phylogenetic trees to determine their relationships with other isolates. Our analysis revealed that the novel OROV genome sequences sampled in 2024 clustered with sequences from the 2022-2024 epidemic into a monophyletic group, with strong bootstrap support (values of 1.0) across all three genomic segments (Figure 3a-c).

• Download figure
• Open in new tab

Figure 3. Molecular Evolution and Demographic History of OROV Segments S, M, and L.

a-c) Maximum likelihood phylogenetic trees of the three OROV segments: S (n = 376), M (n = 231) (a), L (n = 303). Tips are color-coded according to the legend in the left corner; Inner plots indicate the effective population size (i.e., genetic diversity) of OROV infections (Log scale) over time estimated under the coalescent-based Bayesian Skygrid (BSKG) model (posterior median = solid lines, 95% HPD = pale areas) for each segment; d-f) Regression of sequence sampling dates against root-to-tip genetic distances in a maximum likelihood phylogeny of the Brazilian 2022-2024 expansion clade (n⍰=⍰254).

This clade included a basal sequence from the locality of Tefé, Amazonas state, sampled in 2015, indicating a likely Amazonian origin, and has rapidly expanded to other regions, following a north-to-south movement across the country (Figure 3a-c). In addition, this lineage has a reassortant evolutionary history, possessing L and S segments from viruses found in Peru, Colombia, and Ecuador and an M segment from viruses detected in the eastern Amazon region. Detailed analysis of the epidemic expansion, based on estimates of effective population size, revealed a sharp increase at the beginning of 2024 (Figure 3a-c, inner plots), coinciding with the timing of the national surge in OROV cases (Figure 1) (8). In addition, the phylogenetic trees for the S, M, and L segments (Figure 3a-c) revealed two main lineages. The first lineage appeared to be primarily composed of sequences from the northern and northeastern regions, while the second lineage was associated with the southeastern and southern regions. This geographical separation might reflect regional diversification of the virus as it spread outside the Amazon basin.

The observed topological discordance among the phylogenetic trees of different genomic segments indicated the likelihood of multiple reassortment or recombinant events, supporting the notion that reassortment/recombination is a common and potentially significant evolutionary mechanism in bunyaviruses. A recent preprint manuscript provides more information on the potential reassortment events on genomes from the Amazon Basin (9), so this will not be covered in detail in our manuscript.

To reconstruct viral movements across the country, we utilized smaller data sets derived from each genomic segment individually, focusing exclusively on the Brazilian 2022-2024 sublineage (Figure 3 a-c). As there was a strong correlation between sampling time and root-to-tip divergence in all three data sets (Figure 3 g-i), we were able to use molecular clock models to infer evolutionary parameters. Accordingly, we estimated that the mean time of origin for this Brazilian 2022-2024 sublineage to be in early November 2021, with a 95% highest posterior density (HPD) interval ranging from early-August 2021 to early January 2022. This suggests continuous transmission within the Amazon basin (Figure 4a-c). Likely introduced in early 2015, this sublineage remained undetected due to insufficient active surveillance at the national level, initially spreading from the northern part of the country (Amazon basin) and subsequently moving towards the northeastern (Bahia state), midwestern (Mato Grosso state), southeastern (Minas Gerais state), and southern (Santa Catarina state) regions (Figure 4a-c). We expanded the data set related to this Brazilian 2022-2024 sublineage to include international genome sequences recently isolated in Peru and Italy. Analyses of these data revealed that OROV likely initially moved north to south within Brazil and then crossed borders to reach Peru (Figure 4d). Case reports recently detected in Italy are associated with returning travelers from Cuba, although genomes from the latter locality are currently unavailable. It is possible that OROV moved into Cuba sometime in between the events depicted in Figure 4d before being exported into Europe.

• Download figure
• Open in new tab

Figure 4. Inferred Viral Dissemination Patterns of OROV in Brazil and Globally.

a-c) Phylogeographic reconstruction of the spread of OROV (segments S, M, and L) in Brazil. Circles represent nodes of the maximum clade credibility phylogeny and are colored according to their inferred time of occurrence. Shaded areas represent the 80% highest posterior density interval, depicting the uncertainty of the phylogeographic estimates for each node. Solid curved lines denote the links between nodes and the directionality of movement; d) Dissemination patterns of OROV within the Americas and Europe obtained from inferred ancestral-state reconstructions annotated and colored by region. The destination countries of viral exchange routes are shown with dots, with curves indicating the routes from the country of origin to the destination country in a counterclockwise direction. The dashed lines indicate the probable direction of dispersal concluded without evidence due to the paucity of genomes from Cuba.

Discussion

The spread of the novel reassortant OROV lineage between 2022 and 2024 in the Brazilian Amazon region highlights the potential for this virus to emerge and the role of ecological and human factors in its dissemination (11, 12). Our study, along with recent findings (11, 12), underscores the importance of genomic surveillance and phylogenetic analysis in tracing the origin and movement of viral pathogens. The phylogenetic analyses from our study revealed that the OROV sequences sampled between 2022 and 2024 fell into a highly supported monophyletic group, with origins tracing back to a basal sequence from Tefé, Amazonas, Brazil, sampled in 2015. This indicates a likely Amazonian origin for this clade. These findings are consistent with the emergence of a novel reassortant lineage containing the M segment from viruses detected in the eastern Amazon region and the L and S segments from viruses found in Peru, Colombia, and Ecuador (10, 11). This reassortant lineage appears to have emerged in the central region of the Amazonas state between 2010 and 2014, demonstrating a long-range silent dispersion during the latter half of the 2010s.Our study identified approximately 21 reassortment events among the different genome segments, supporting the role of reassortment in the evolutionary history of OROV (11). It is possible that this contributed to the virus’s ability to adapt to new ecological niches and hosts, fueling its spread across different regions of Brazil. The sharp increase in OROV cases in different Brazilian regions along with the concurrent rise in real-time RT-PCR testing, reflects heightened surveillance and detection efforts underscoring the widespread concern and the proactive measures taken to monitor this emerging pathogen.

Our genome-based surveillance approach revealed a clear north-to-south movement of the virus within the country. The peak of OROV transmissions coincided with the rainy season in the Amazon basin, suggesting that environmental factors may play a major role in the virus’s transmission dynamics. Deforestation and climate change have contributed to the spread of OROV beyond the Amazon basin (32, 33). The alteration of ecosystems through deforestation disrupts natural habitats and promotes the migration of vectors, such as midges, into new areas (33). Climate change exacerbates this by altering weather patterns and creating favorable conditions for vector proliferation (34). These ecological changes facilitate the movement of OROV from its traditional Amazonian confines to other regions of Brazil and beyond. Recent studies have shown that increased deforestation rates and rising temperatures in the Amazon basin correlate with the expanded range of vector species capable of transmitting OROV (11, 35-40). This north-to-south dissemination pattern highlights the importance of addressing environmental and climate-related factors in controlling the spread of arboviruses. The prolonged cryptic circulation of the virus emphasizes the critical need for robust, active screening programs to monitor and control the spread of such pathogens effectively. The detection of three fatal cases associated with Oropouche virus (OROV) infection, even in patients without reported comorbidities, suggests that the clinical impact of OROV might be more significant than previously understood. This necessitates further investigations into adverse pregnancy outcomes and the potential for vertical transmission of OROV. On August 3rd, the Brazilian Ministry of Health (BrMoH) confirmed the first fetal death due to OROV with mother-to-child transmission in the state Pernambuco, located in the northeastern part of the country (41). This highlights the need for further research into the factors contributing to severe disease outcomes and the development of effective treatment strategies. In addition, the identification of long-range OROV migrations facilitated by human activities, as well as viral expansion beyond Brazilian borders to Peru, Cuba, and subsequently Europe, underscores the global implications of this emerging pathogen. The role of climate change in the expansion of vectors into new regions, coupled with increased human mobility, necessitates a coordinated international response to monitor and control the spread of OROV and other arboviruses.

Our study, in conjunction with recent research, emphasizes the critical need for continuous and widespread genomic surveillance. By understanding the evolutionary and epidemiological patterns of OROV, we can better anticipate and mitigate future outbreaks, ensuring more effective public health responses both within Brazil and globally. This approach will be instrumental in managing the ongoing spread of OROV and preparing for potential future threats posed by similar viral pathogens.

Author Contributions

Conceptualization: V.F., J.L., T.d.O., L.C.J.A., and M.G.; Methodology: F.C.M.I., F.M.P., E.C.O., J.T.N.R., M.H.M., V.F., T.E.R.A., N.R.G., L.M.R.T., M.K.A.G., V.B.M., A.A.R., A.R., A.G.F., A.L.S.M., B.M.N.F., C.F.C.A., D.S.P., F.G.M.L., F.V.M.S., G.C.P., H.T., J.D.,P.,C.A., K.M.F.M., K.R.V., L.C.S., L.C.M.S., L.C.V.F., L.O.M., M.A.C., M.S.O., M.P.D.A., M.C., M.T.L., M.A.P., M.L.C.R., P.E.S.S., P.R., R.S.A., R.S., R.T.G., R.V.C., S.G., S.C.F.S., S.A.A.B., S.H.S.P.P., S.I.R.A., S.F.S., W.V.V., D.P.M., E.C.H., T.d.O., J.L., L.C.J.A., and M.G.; Investigation: F.C.M.I., F.M.P., E.C.O., J.T.N.R., M.H.M., V.F., T.E.R.A., N.R.G., L.M.R.T., M.K.A.G., V.B.M., A.A.R., A.R., A.G.F., A.L.S.M., B.M.N.F., C.F.C.A., D.S.P., F.G.M.L., F.V.M.S., G.C.P., H.T., J.D.,P.,C.A., K.M.F.M., K.R.V., L.C.S., L.C.M.S., L.C.V.F., L.O.M., M.A.C., M.S.O., M.P.D.A., M.C., M.T.L., M.A.P., M.L.C.R., P.E.S.S., P.R., R.S.A., R.S., R.T.G., R.V.C., S.G., S.C.F.S., S.A.A.B., S.H.S.P.P., S.I.R.A., S.F.S., W.V.V., D.P.M., E.C.H., T.d.O., J.L., L.C.J.A., and M.G.; Data curation: V.F., J.L., and M.G.; Original draft preparation: J.L., M.G.; Review and editing: F.C.M.I., F.M.P., E.C.O., J.T.N.R., M.H.M., V.F., T.E.R.A., N.R.G., L.M.R.T., M.K.A.G., V.B.M., A.A.R., A.R., A.G.F., A.L.S.M., B.M.N.F., C.F.C.A., D.S.P., F.G.M.L., F.V.M.S., G.C.P., H.T., J.D.,P.,C.A., K.M.F.M., K.R.V., L.C.S., L.C.M.S., L.C.V.F., L.O.M., M.A.C., M.S.O., M.P.D.A., M.C., M.T.L., M.A.P., M.L.C.R., P.E.S.S., P.R., R.S.A., R.S., R.T.G., R.V.C., S.G., S.C.F.S., S.A.A.B., S.H.S.P.P., S.I.R.A., S.F.S., W.V.V., D.P.M., E.C.H., T.d.O., J.L., L.C.J.A., and M.G.; Visualization: V.F., J.L., and M.G. All authors have read and agreed to the published version of the manuscript.

Conflicts of Interest

The authors declare no conflict of interest.

Supplementary figures

• Download figure
• Open in new tab

Figure S1. Symptoms versus age and gender.

The (left) age distribution of cases reported with different symptoms and (b) proportion of cases reporting different symptoms, both disaggregated by gender. Data includes the metadata for the sequenced samples.

• Download figure
• Open in new tab

Figure S2. Sequence Coverage Probability Across Genome Segments in Different Brazilian States.

The figure displays the probability of coverage for all sites across four different genome segments (S, M, L, and the entire genome) in relation to the CT value. The probability was estimated using a Generalized Additive Model (GAM; as done previously e.g. for CHIKV sequencing data [Giovanetti et al Emerg Infect Dis. 2023;29(9):1859-1863]). The GAM was defined with a binomial family, with coverage modeled as the proportion of sites recovered from sequencing, using sample state (Brazilian state) as random effect. Solutions presented (in color) include the specific ones for the two states with most sequencing (Bahia, red; Minas Gerais, blue) and the general model output (gray).