Genomic characterization, origin, and local transmission of Oropouche virus in Bolivia in 2024

Summary

Background

The Oropouche virus (OROV) is an arthropod-borne virus that causes an acute febrile illness, like other arboviral diseases. In 2024, Oropouche cases sharply increased in several countries of the Americas, including Bolivia. Here, we investigate the origin and spread of OROV in the Bolivian Amazon region.

Methods

Full-length OROV genomes from 34 positive samples collected in the three affected Bolivian departments during the 2024 outbreak were sequenced using an amplicon-based approach. Maximum Likelihood (ML) phylogenetic analyses of separate viral segments were conducted to identify the responsible viral lineage. Bayesian phylogeographic analysis of concatenated viral segments was used to reconstruct the viral spatiotemporal dispersion pattern within the country.

Findings

The first Oropouche cases in Bolivia 2024 were reported using samples collected from the Pando department during mid-January, and the peak of Oropouche cases occurred in mid-April. The phylogenetic analysis of OROV genomes revealed that all cases detected in Bolivia belong to the novel reassortant OROV clade that drove the recent epidemic in Brazil. Our phylogeographic analysis detected at least two exportation events from the Brazilian state of Acre to the Bolivian municipalities of Guayaramerín and Riberalta, both located in the Beni department, with subsequent dissemination to municipalities of Pando and La Paz departments. Viral introductions likely occurred between early October and early November 2023, indicating a lag of approximately three months between the introduction of OROV and its detection.

Interpretation

Our findings confirm that OROV spread at least twice from the western Brazilian Amazon to the neighboring Bolivian department of Beni in late 2023, successfully establishing regional transmission chains. These findings underscore the critical need for active OROV surveillance across the border Amazonian region between Brazil and Bolivia. They also confirm the potential for sustained OROV transmission within the Bolivian Amazon, highlighting the importance of preparedness for future outbreaks.

Funding

This publication was in part supported by the Cooperative Agreement Number NU50CK000639 awarded to the Pan American Health Organization and funded by the Centers for Disease Control and Prevention.

Research in context.

Evidence before this study

Before 2024, large outbreaks of Oropouche virus (OROV) were predominantly reported in the Amazon regions of Brazil and Peru. However, in 2024, significant outbreaks first emerged in the Brazilian Amazon region. They were soon followed by a surge of cases in the neighboring South American countries of Bolivia, Colombia, and Peru. We searched PubMed and preprint servers (medRxiv and bioRxiv) available as of October 25, 2024, for studies examining the circulation of OROV in Bolivia, using the terms [“Oropouche” AND “Bolivia”]. We identified only one study that reported a few anecdotal cases of past OROV infections in Bolivia, relying on serological tests and a couple of reviews. Therefore, it was important to properly investigate how OROV emerged in Bolivia in 2024.

Added value of this study

This is the first study to analyze the genomic characteristics of OROV circulating in Bolivia. In this study, we sequenced 34 full-length OROV genomes, representing 10% of all RT-qPCR-confirmed OROV cases at that time across Pando, Beni, and La Paz departments between January and May 2024. The OROV detected in Bolivia belongs to the novel reassortant lineage recently identified in Brazil. We identified at least two introductions of OROV from the western Brazilian Amazon region into the neighboring Bolivian department of Beni around late 2023, followed by its spread to other regions within Bolivia during the rainy season. Our estimates indicate that the virus circulated in Bolivia for approximately three months before the first case was detected.

Implications of all the available evidence

Our study confirms that the novel OROV reassortant lineage recently identified in Brazil rapidly disseminated across the Amazonian border into Bolivia. The successful establishment of OROV in Bolivia indicates that the country possesses suitable ecological conditions to support sustained transmissions of this arbovirus. Our findings also emphasise the crucial need for active and sustained molecular surveillance of OROV in the Bolivian Amazon region to enable the timely detection of new outbreaks in the country.

Introduction

The Oropouche virus (OROV) is an arthropod-borne virus primarily transmitted by the bite of the hematophagous midge Culicoides paraensis.¹ The symptomatic infection with this virus causes Oropouche fever, an acute febrile illness that may be confused with dengue or Chikungunya in the early phase. However, rare but more severe manifestations, including neuroinvasive diseases, have been known for decades.¹ During the 2023–2025 outbreak, maternal-fetal complications and fatal outcomes were reported for the first time.2, 3, 4 OROV (Orthobunyavirus oropoucheense) belongs to the Simbu serogroup within the Orthobunyavirus genus of the Peribunyaviridae family; this virus has a negative-sense, single-stranded RNA genome composed of large (L), medium (M), and small (S) segments.⁵

OROV was first detected in Trinidad and Tobago in 1955.⁶ Most outbreaks until 2000 were reported in Brazil and Peru, and over the past 25 years, sporadic cases of OROV have been identified in other American countries, including Argentina, Bolivia, Colombia, Ecuador, French Guiana, Haiti, and Panama.⁷ In 2024, the number of OROV cases sharply increased in the Americas, with 10,183 confirmed cases as of the epidemiological week 40 of 2024. Autochthonous Oropouche fever cases have been reported across six countries: Brazil (n = 8258), Peru (n = 936), Cuba (n = 555), Bolivia (n = 356), Colombia (n = 74), Ecuador (n = 2), and Guyana (n = 2).⁸ The recent expansion of the virus outside the Amazon basin and the broad distribution of its primary vector have raised serious concerns about the potential spread of OROV beyond its historical range.

The presence of OROV in Bolivia was first detected in Cochabamba as part of a clinical surveillance program investigating febrile disease etiologies from 2000 to 2007.⁹ However, this study analyzed a limited number of locations and only used serological tests for indirect OROV detection. Systematic molecular surveillance of OROV was initiated by the Ministry of Health and Sports in Bolivia in 2024, following the epidemiological alert for OROV in the Americas issued by the Pan American Health Organization (PAHO).¹⁰ Two laboratories under the auspicious of Bolivia Ministry of Health and Sports reported a total of 356 OROV positive cases; these cases were confirmed by RT-qPCR between January and May 2024, in both rural and urban areas across the departments of Pando, Beni, and La Paz, all located within Bolivia's Amazon basin.¹¹ This study aimed to elucidate the origin and spatiotemporal dispersion pattern of the virus responsible for the current Oropouche fever outbreak in Bolivia. To this end, we comprehensively analyzed genomic sequences derived from 34 OROV-positive samples collected across the three affected departments (Pando, Beni, and La Paz) in 2024.

Methods

OROV positive samples and epidemiological data

In Bolivia, between January and May 2024 (epidemiological weeks 1–22), the Ministry of Health and Sports reported 356 positive cases of the OROV, which were detected by the National Center for Tropical Diseases (CENETROP) and by the National Institute of Health Laboratories (INLASA). CENETROP received 1130 serum samples from patients suspected of arbovirus infection sent from La Paz, Beni, Pando, and Santa Cruz departments. Of the 1130 serum samples analyzed, 974 were negative, and 156 were positive for detecting OROV using the duplex real-time PCR assay developed by Naveca et al. , which PAHO has recommended for the molecular surveillance of Oropouche and Mayaro viruses.¹²

Whole-genome sequencing and genome assembling

Among the 156 positive samples identified by CENETROP between January and May 2024, 34 samples (22%) were selected for whole genome sequencing based on CT values (<25), epidemiological week (EW), and spatial location. Each selected sample had an associated identification code and associated epidemiological metadata such as EW, spatial location (Department and Municipality), biome, age, sex, hospitalization, date of symptoms onset, collection date, and date of notification (Supplementary Table S1).

Complete genome sequencing was carried out with an amplicon-based strategy previously developed by the Emergent, Reemergent, or Neglected Viruses Surveillance Laboratory (ViVER) at Fiocruz, Amazonas, Brazil.¹³ Library preparation was carried out with Illumina's COVIDSeq kit, and sequencing was carried out on the MiniSeq version 2.3.0 sequencer with the Mid-Output Reagent Cartridge kit (300 cycles) for a 151 bp x 2 pair-end run. FASTQ reads were generated at Illumina BaseSpace (https://basespace.illumina.com) and imported into Geneious Prime v2024.0.5. Initially, reads were trimmed for quality, duplicates were removed, and the remaining reads were normalized and assembled into contigs using a custom workflow that employs the tools BBDuk, Dedupe, BBNorm, and BBMap (v.39.06), embedded into Geneious Prime. The sequences AF484424, AF441119, and AY237111 from GenBank were used as reference sequences for the S, M, and L virus segments of OROV, respectively. The consensus sequences were extracted using a threshold of at least 50% to call a base.

OROV whole-genome genotyping

The 34 complete OROV sequences of L, M, and S genomic segments generated here were first aligned with corresponding segments of published full-length (≥70% of coverage) OROV genome sequences available at the National Center for Biotechnology Information (NCBI), sampled in the Americas between 1955 and 2024, containing the prototype sequences of OROV (L: AF484424; M: AF441119; S: AY237111), Iquitos virus (L: KF697142; M: KF697143; S: KF697144), Perdões virus (KP691627; M: KP691628; S: KP691629), and Madre de Dios virus (L: KF697147; M: KF697145; S: KF697146), as well as OROV sequences representing the major M (M1 and M2), L (L1 and L2) and S (S1, S2 and S3) clades described by Naveca et al., 2024 (Supplementary Table S2).¹³ Next, an alignment was generated comprising the 34 complete OROV sequences from Bolivia and all sequences from the OROV_{BR-2015–2024} clade, including the sub-clades AMACRO-I, AMACRO-II, (AMACRO is a Brazilian region that comprises the states of Acre, Rondônia and the southern part of Amazonas state), AM-I, AM-II, AM-III, and RR-I, as described by Naveca et al.¹³ This alignment was performed separately for each of the three genome segments (L, M, and S) (Supplementary Table S3) as well as for the concatenated segments (Supplementary Table S4). Sequences were aligned with MAFFT v7.490,¹⁴ and maximum likelihood (ML) phylogenetic trees were inferred with IQ-TREE v2.1.1¹⁵ using the best substitution model defined by the software ModelFinder.¹⁶ Branch support was inferred using ultra-fast bootstrap on 1000 replicates. The ML trees were visualized using FigTree v.1.4.4 (https://github.com/rambaut/figtree).

Bayesian evolutionary and phylogeographic analyses

To reconstruct the spatiotemporal history of OROV spread in Bolivia, we selected a subset of near-full-length sequences of the OROV_BR-2015-2024 epidemic clade that includes the most basal sequence sampled in the Amazonas state in 2015, plus all sequences belonging to the sub-clades AMACRO-I and AMACRO-II (which comprises all sequences sampled in Bolivia, as observed in the ML inference) (Supplementary Table S5). The resulting dataset of concatenated segments L, M, and S was submitted to ML phylogenetic reconstruction as described above, and the temporal structure was estimated by performing a root-to-tip linear regression with TempEst v.1.5.3.¹⁷ Time-scaled phylogeographic trees for concatenated genomic segments were estimated using the Bayesian Markov chain Monte Carlo (MCMC) approach, implemented in the software BEAST v1.10.¹⁸ Time-scaled trees were inferred with a relaxed uncorrelated lognormal distributed molecular clock model with a continuous-time Markov chain (CTMC) rate reference prior¹⁹ and the non-parametric Bayesian skyline coalescent demographic model.²⁰ The ancestral node states were reconstructed with a CTMC prior, and a discrete spatial diffusion with a symmetric substitution model complemented with a Bayesian stochastic search variable selection (BSSVS) to identify the significant migration routes.²¹ Markov Chain Monte Carlo (MCMC) was run for 100 million generations, and convergence was assessed by calculating the Effective Sample Size (ESS) for all parameters using Tracer v1.7.1.²² The maximum clade credibility (MCC) trees were summarized with TreeAnnotator v.1.10 and visualized using FigTree v.1.4.4.¹⁸

Statistical analysis

The significance of the correlation between the collection date and root-to-tip genetic divergence was assessed with a Spearman correlation test. The threshold for statistical significance was set to P < 0.05. Statistical analyses were performed using GraphPad Prism version 9.0 software.

Ethics statement

This study was approved by the Ethics Research Committee of the Universidad Cristiana de Bolivia (FWA00028929) under protocol number CEI/001/2024, which waived signed informed consent.

Role of the funding source

The funders of the study had no role in study design, data collection, data analysis, data interpretation, or writing of the manuscript.

Results

Epidemiological data of Oropouche fever cases in Bolivia

As of EW 22 of 2024, the Ministry of Health of Bolivia notified 356 Oropouche cases confirmed by real-time RT-PCR at the national level, of which 156 cases were reported by the CENETROP laboratory (Fig. 1a). The first OROV-positive cases were reported in samples collected from the Pando department during EW 3 of 2024, and the peak of cases occurred in EW 15, coinciding with the Amazon region's rainy season (from November to April). Oropouche cases have decreased after EW 16, with the last cases reported in EW 20. Most cases were identified in the La Paz department (75%, 268/356), followed by Beni (21%, 76/356) and Pando (3%, 12/356) (Fig. 1b). The cases were reported among 16 municipalities, with the highest proportion of cases reported in the municipalities of Irupana (33%, 117/356), La Asunta (13%, 45/356), and Chulumani (12%, 44/356) in the La Paz department, and in the municipality of Guayaramerín (12%, 44/356) in the Beni department. Oropouche cases were evenly distributed among males and females (Table 1, detailed demographics and clinical information at Supplementary Table S6). The highest proportion of Oropouche cases occurred in the <19 age group for males (28.5%, 49/172), and in the 30–39 age group for females (24.7%, 44/178) (Fig. 1c). From the 156 positive cases of Oropouche detected in CENETROP, we recovered symptom information for 147 (94.2%). The most common symptoms reported for OROV infections were fever (96.6%, 142/147), headache (93.9%, 138/147) and myalgia (83.0%, 122/147), followed by nausea (40.8%, 60/147), vomiting (27.2%, 40/147), loss of appetite (11.6%, 17/147) and photophobia (9.5%, 14/147) (Fig. 1d).

Fig. 1.

The 2024 Oropouche virus outbreak in Bolivia. (a) Temporal distribution of Oropouche cases and sampled genomes across Bolivian departments, with the corresponding map on the panel's right side. The graphs and the map are color-coded according to the legend at the bottom of the panel. (b) Spatial distribution of Oropouche cases and sampled genomes across Bolivian cities. In both maps, the circle's diameter is proportional to the number of cases and genomes, as indicated by the respective legends. Only cities with more than five genomes are labelled in the first map, while in the second map, all cities with sampled genomes are annotated. (c) Proportion of OROV-positive cases in Bolivia, categorized by age and sex. (d) The proportion of symptoms recorded from OROV-positive cases in Bolivia is stratified by sex. AR: Argentina, BO: Bolivia, BR: Brazil, CL: Chile, PE: Peru, UY: Uruguay.

Table 1.

Characteristics of the population with confirmed Oropouche virus infection.

Patients	Overall, n (%)	Female, n (%)	Male, n (%)
Demographics
Sex
	356 (100)	179 (50.3)	177 (49.7)
Symptoms
W.S.I.	147 (41.3)	80 (44.7)	67 (37.8)
M.S.I.	209 (58.7)	99 (55.3)	110 (62.2)
Agea
<19	89 (25.4)	40 (22.5)	49 (28.5)
20–29	60 (17.1)	30 (16.9)	30 (17.4)
30–39	70 (20.0)	44 (24.7)	26 (15.1)
40–49	54 (15.4)	26 (14.6)	28 (16.3)
>50	77 (22.0)	38 (21.3)	39 (22.7)
Detailed symptomsb
Fever	142 (96.6)	65 (97.0)	77 (96.3)
Headache	138 (93.9)	61 (91.0)	77 (96.3)
Myalgia	122 (83.0)	58 (86.6)	64 (80.0)
Nausea	60 (40.8)	67 (34.3)	37 (46.3)
Vomiting	40 (27.2)	9 (13.4)	31 (38.8)
Loss of appetite	17 (11.6)	8 (11.9)	9 (11.3)
Photophobia	14 (9.5)	8 (11.9)	6 (7.5)
Dizziness	9 (6.1)	6 (9.0)	3 (3.8)
Abdominal pain	8 (5.4)	3 (4.5)	5 (6.3)
Walking difficulties	8 (5.4)	2 (3.0)	6 (7.5)
Artralgia	3 (2.0)	2 (3.0)	1 (1.3)
Mucosal bleeding	1 (0.7)	0 (0.0)	1 (1.3)
Petechiae	1 (0.7)	1 (1.5)	0 (0.0)

W.S.I. = with symptoms information; M.S.I. = missing symptoms information.

^aSix cases have missing age information.

^bDetailed symptoms information for 147 patients.

Spatiotemporal patterns of the OROV spread across Bolivia

The 34 full-length OROV genomes generated in this study from samples collected between EW3 and EW20 are from all three Bolivian departments (Fig. 1a and b). The ML phylogenetic analyses performed individually for each genomic segment revealed that all 34 OROV sequences belong to the novel reassortant OROV_BR-2015-2024 clade that drove the recent epidemic in Brazil (Fig. 2).¹³ A closer inspection of the tree topology indicates that the OROV_BR-2015-2024 clade circulating in the western Amazon evolved without major reassortment events (Supplementary Fig. S1). Consequently, to maximize phylogenetic resolution, all subsequent analyses focusing exclusively on sequences from the OROV_BR-2015-2024 clade were conducted using concatenated genomic segments.

Fig. 2.

Maximum Likelihood phylogenetic analyses of the M, L and S segments of OROV. Phylogenetic trees inferred from the segments M (a), L (b), and S (c) of OROV sequences with complete genomes (n = 131). The tips of Bolivian viruses are colored in light brown with an increased diameter, and prototypical OROV sequences in darker brown. Brackets demarcate major OROV clades alongside their denomination and statistical support (aLRT). All trees are drawn according to the genetic distance scale at the bottom of each panel.

The ML phylogenetic analysis of the OROV_BR-2015-2024 clade performed with concatenated segments L, M, and S indicates that OROV sequences from Bolivia are nested within a Brazilian sub-clade previously named AMACRO-II, which is one of the major sub-clades identified in the Brazilian Amazonian region (Fig. 3). According to a previous study, the sub-clade AMACRO-II most probably emerged in the Acre state in August 2023 (2023-08-24, 95% Highest Posterior Density [95% HPD]: 2023-06-26 to 2023-10-14) and comprised all OROV sequences detected in the Brazilian states of Acre and Rondônia, which borders Bolivia, and all sequences from the southern region of the Amazonas state sampled from December 2023 to November 2024.¹³ OROV sequences from the AMACRO region (Southern Amazonas, Acre, and Rondônia) and Bolivia form a distinct cluster separate from those sampled in other regions of Amazonas state (sub-clades AM-I, AM-II, and AM-III) from December 2023 onward, as well as from sequences recovered in Roraima state (sub-clade RR-I) between 2022 and 2024.

Fig. 3.

Maximum Likelihood phylogenetic tree of concatenated segments of sequences belonging to the OROVBR-2015–2024 clade. Major OROVBR-20125-2024 sub-clades are highlighted with their respective designations and annotated statistical support (aLRT). Tip points are color-coded according to their sampling locations, which are also reflected on the map displayed on the left side of the panel. Bolivian sequences are marked with slightly larger shapes for easier identification. To improve visualization, the sub-clades from the Brazilian state of Amazonas are collapsed into triangles. The tree is drawn according to the scale at the bottom of the left panel. AM: Amazonas, BO: Bolivia, BE: Beni, BR: Brazil, CO: Colômbia, EC: Ecuador, LP: La Paz, PE: Peru, PN: Pando, PY: Paraguay, RO: Rondônia, RR: Roraima.

Bayesian phylogeographic analysis with concatenated L, M, and S segments was then performed to model the viral diffusion process between Brazil and Bolivia in a discrete space. For this analysis, we retained the oldest sequence of OROV_BR-2015-2024 clade detected in the Amazonas state in 2015, and all OROV sequences belonging to sub-clades AMACRO-I (detected between January and June 2023) and AMACRO-II (detected from December 2023 onwards). OROV sequences from Bolivia, Acre, Amazonas, and Rondônia were grouped by municipality of origin. None of the patients in the present study reported having a travel history. The correlation between genetic divergence and sampling time was significant for this OROV dataset (P < 0.05) (Fig. 4a), supporting a robust temporal structure. The genome-wide molecular clock rate of the dataset was estimated at 1.2 × 10⁻³ (95% HPD: 1.0 − 1.5 × 10⁻³) substitutions/site/year, similar to that previously estimated for the complete dataset of OROV_BR-2015-2024 clade.¹³

Fig. 4.

Spatio-temporal dynamics of the OROV spread in Bolivia. (a) Linear regression of root-to-tip genetic divergence against sampling dates for concatenated segments of sequences belonging to the OROVBR-2015-2024 sub-clades AMACRO-I, AMACRO-II, and the oldest sequence sampled in AM-Tefe in 2015 (n = 92). Samples are color-coded by department/state, as per the color scheme shown on the right side of the panel. (b) Time-scaled MCC tree of the concatenated segments of sequences belonging to the OROVBR-2015-2024 sub-clades AMACRO-I, AMACRO-II, and the oldest sequence sampled in AM-Tefe in 2015 (n = 92). Tips are color-coded by sampling location (municipality), and branches are colored according to inferred ancestral locations, following the scheme on the bottom right of the panel. Bolivian clusters BO-I and BO-II are annotated in the tree. The sequence sampled in AM-Tefe in 2015 was excluded for visual clarity. (c) Map of Bolivia highlighting the sub-national departments. Lines represent inferred viral migrations as reconstructed through discrete phylogeographic analysis, connecting origins and destinations of migrations with arcs oriented clockwise and color-coded according to the Bolivian cluster, as indicated in the legend at the bottom of the panel. (d) The median TMRCA (circle) and its 95% HPD interval (transparent polygon) of the BO-I and BO-II subclusters are shown, along with the period of cryptic circulation (thin line) and the sampling range (thick line). AC: Acre, AM: Amazonas, BN: Beni, BR: Brazil, CL: Chile, CO: Colômbia, EC: Ecuador, RO: Rondônia, LP: La Paz, PE: Peru, PN: Pando, PY: Paraguay.

Bayesian phylogeographic analysis revealed that the 34 OROV sequences from Bolivia formed two monophyletic sub-clades, designated BO-I (n = 22) and BO-II (n = 12) (Fig. 4b). Rio Branco, the capital of Acre state, was identified as the most probable source of the Bolivian sub-clades BO-I (posterior state probability [PSP] = 0.79) and BO-II (PSP = 1) (Fig. 4c). The BO-I and BO-II sub-clades were likely introduced into the nearby municipalities of Guayaramerín (PSP = 0.92) on October 12, 2023 (95% HPD: September 5–November 17, 2023) and Riberalta (PSP = 0.84) on November 10, 2023 (95% HPD: September 24–December 31, 2023), respectively, both located in the Beni department (Fig. 4d). From Guayaramerín, the BO-I sub-clade spread westward to the municipality of Cobija (Pando department) and southward to Teoponte (La Paz department) (Fig. 4c). The BO-I sub-clade further spread from Cobija to San Buenaventura (La Paz department), from Teoponte to Palos Blancos (La Paz department), and from Palos Blancos to San Borja (Beni department) (Fig. 4c). From Riberalta, the BO-II sub-clade spread eastward to the municipality of Guayaramerín and southward to San Buenaventura and Irupana (La Paz department) (Fig. 4c). The lag between the median time to the most recent common ancestor (T_MRCA) and the earliest genome sequence for each Bolivian sub-clade was 99 days for BO-I and 82 days for BO-II (Fig. 4d).

Discussion

In 2024, Bolivia experienced its largest documented Oropouche fever outbreak to date, affecting 16 municipalities in three departments of the Amazon basin (Pando, Beni, and La Paz). Interestingly, during this outbreak, no cases of Oropouche were reported in the department of Cochabamba, where the first cases of Oropouche were reported between 2005 and 2007.⁸ The 2024 OROV epidemic in Bolivia spanned from January to May and peaked in mid-April, thus coinciding with the rainy season (November to April). The same phenomenon was observed in the Brazilian Amazon Region, where the increase in OROV circulation coincides with the rainy season (December to March).¹³ This association is likely attributable to the expansion of humid habitats, which create favorable conditions for the development of C. paraensis larvae. Previous studies have confirmed the presence of this vector in the departments of Cochabamba and La Paz.^23,24 However, to date, no published reports document its occurrence in Beni and Pando. Further studies are required to confirm the correlation between the onset of the rainy season, the increased presence of midges, and the occurrence of OROV outbreaks in Bolivia's Amazon region.

Most Oropouche cases during the 2024 epidemic in Bolivia were mild, and the most common clinical manifestations (fever, headache, and myalgia) were the same as those observed in both previous and ongoing (2023–2025) OROV epidemics.^13,25 A key difference was the very low incidence of arthralgia (<2%) and the complete absence of retro-orbital pain in Bolivia, both of which were reported at much higher frequencies (30–40%) in previous outbreaks and in other locations.^13,25 We did not observe severe Oropouche fever cases during this study in Bolivia. However, the first two deaths linked to OROV infection and several cases of maternal-fetal complications,^2,4 all first reported in Brazil, highlight the importance of strengthening OROV surveillance across the Americas.

Phylogenetic analyses confirm that all OROV cases detected in Bolivia belong to the OROV_BR-2015-2024 clade, which originated from a new reassortment event around 2010–2015 and that was responsible for the increase of OROV cases in the Brazilian western Amazon region between 2022 and 2024.¹³ These analyses further revealed that the OROV sequences from Bolivia were part of the AMACRO-II sub-clade, which circulates in the Brazilian states of Acre and Rondônia since late 2023,¹³ two states that border the departments of Pando and Beni, respectively, where the first Oropouche cases were detected in Bolivia. The discrete phylogeographic analysis indicates that the AMACRO-II sub-clade spread from Rio Branco, the capital of Acre state, to the municipalities of Guayaramerín and Riberalta in the Beni department and to the Brazilian state of Rondônia. Thus, the OROV AMACRO-II sub-clade was probably repeatedly transmitted across the neighboring Amazonian regions of Acre, Rondônia, and Beni in the bordering region between Brazil and Bolivia.

Notably, the two most probable entrance points of OROV in Bolivia, the municipalities of Guayaramerín and Riberalta, are locations closer to Rondônia state than Acre state. Particularly, the Bolivian city of Guayaramerín is located on the left bank of the Mamoré River, in front of the Rondônia city of Guajará-Mirim, in Brazil. These twin cities display an intense and dynamic daily trade flow, potentially making them a critical area for the transborder spread of OROV.²⁶ Despite the geographical proximity, our discrete phylogeographic analysis failed to detect direct viral migrations from Rondônia to Bolivia. However, Rondônia experienced a significant OROV epidemic in 2023–2024, and few viral sequences from late 2023 to early 2024 from this state have been described. Consequently, most Brazilian sequences within the AMACRO-II sub-clade sampled so far were from the state of Acre. We hypothesize that this geographic sampling bias may have influenced the accuracy of our phylogeographic inference, and the inclusion of a larger number of OROV sequences from Rondônia may enhance the precision of this reconstruction.

Our analysis suggests that the BO-I subclade was likely introduced in Guayaramerín around October 2023, while the BO-II subclade was introduced in Riberalta around November 2023. Both OROV subclades spread successfully beyond the Beni department, reaching the departments of Pando (BO-I) and La Paz (BO-I and BO-II) within the Amazon basin. Therefore, we estimate that OROV circulated undetected for approximately 2–3 months before its first identification in Bolivia in January 2024. Our results underscore the urgent need to enhance molecular surveillance of OROV in cross-border areas and confirm OROV's capacity for rapid spread throughout the Bolivian Amazon region, which constitutes approximately 65% of Bolivia's territory. Intensive and regular testing of individuals presenting symptoms consistent with OROV infection in border Bolivian localities, such as Guayaramerín (near Rondônia) and Cobija (near Acre), could support the establishment of an early warning system to shorten the period of undetected viral spread. Importantly, OROV also spread outside the Brazilian Amazon region, which suggests the importance of strengthening the differential diagnosis of OROV and entomological surveillance for the OROV vector in Bolivian departments beyond the Amazon basin, like south of La Paz, Potosí, Oruro, south of Cochabamba and Tarija.²⁷

The precise causes of the recent upsurge of OROV in Bolivia are unclear. First, estimating the true prevalence of OROV infection in Bolivia during previous decades is challenging, as systematic surveillance for the virus was not in place before 2024. However, a recent study based on in-vitro data suggested that the novel reassortant OROV lineage could lead to less neutralization by antibodies, more efficient infection of vectors, and greater virulence against previous OROV strains.²⁸ However, conflicting findings from a subsequent study indicate that the phenotypic characteristics of the emergent OROV lineage, initially described by Naveca et al., require further evaluation.²⁹ Moreover, the recent spread of a novel reassortant OROV in the Amazon region coincided with a remarkable El Niño Southern Oscillation (ENSO) event, which caused above-average temperatures (maximum and minimum) in all Bolivian departments in the period from November 2023 to March 2024.¹³ This climatological factor, combined with agricultural expansion and deforestation, may have contributed to the transmission of OROV in Bolivia.³⁰

The major limitation of our study is the geographic-based sampling bias due to the low number or absence of sequence data from some key locations in Bolivia and the neighboring Brazilian states of Acre and Rondônia, which may have impacted the accuracy of phylogeographic inferences presented in this study. For example, although the municipalities of La Asunta (13%) and Chulumani (12%) in the La Paz department account for a significant proportion of OROV cases in Bolivia in 2024, no viral genomes have been recovered from these locations. Another important limitation is the temporal sampling bias. Most OROV genomes from Rondônia (91%) were recovered during the first half of 2023, whereas most viral genomes from Acre (77%) were obtained between late 2023 and early 2024. Consequently, despite detecting four times as many OROV-positive cases in Rondônia (n = 1714) compared to Acre (n = 396) between November 2023 and May 2024, the number of OROV genomes recovered in Rondônia (n = 2) was substantially lower compared to Acre (n = 20) during the same period.

In conclusion, our study confirms that the novel Brazilian OROV_BR-2015-2024 clade was introduced at least two times and cryptically circulated for 2–3 months in Bolivia. The spatiotemporal pattern supports that OROV dissemination was initially driven by short-distance movements from the Brazilian states of Acre or Rondônia to the neighboring Bolivian department of Beni and subsequent spread from Beni to other Bolivian departments. These findings demonstrate the epidemic potential of the new reassortant OROV lineage to spread beyond Brazil's borders and underscore the importance of maintaining active surveillance for emerging and reemerging viruses, particularly in cross-border regions of the Amazon shared by Bolivia and Brazil.

Contributors

JACV, SSM, COM, and LRLM did the genomic sequencing and had access to raw data. JACV and FGN verified the data. JACV, IA, FGN, SSM, GB, and LRLM performed the phylogenetic analyses. JACV, GB, IA, SSM, LG, MMG, FGN, and LRLM wrote the article. HCL, RSK, and MXEM assisted with the epidemiological analysis. JACV, IA, SSM, COM, NML, JDM, LG, MMG, JL, LF, JMR, GB, FGN, and LRLM contributed to the overall design, reviewed, and commented on draft articles. LG, MMG, JL, LF, and JMR were responsible for funding acquisition. JACV, LRLM, JMR, GB, and FGN had final responsibility for the decision to submit for publication.

Data sharing statement

All OROV consensus sequences generated in this study (n = 102, corresponding to the three segments of 34 unique samples) were deposited in GenBank (https://www.ncbi.nml.nih.gov/genbank) under the accession numbers PQ634468–PQ634569. The BEAST xml file is available at https://github.com/larboh-ioc/orov_bo.

Editor note

The Lancet Group takes a neutral position with respect to territorial claims in published maps and institutional affiliations.

Declaration of interests

FGN declares receiving grants from CNPq and FAPEAM, Brazil; support for attending meetings from PAHO and Brazilian MoH; A patent issued by Brazil INPI for a device for LAMP nucleic acid amplification; and participation as a member of the Arboviruses advisory group of the Brazilian Ministry of Health. The other authors declare that they have no competing interests.