Air Travel Is Associated with Intracontinental Spread of Dengue Virus Serotypes 1–3 in Brazil

Dengue virus and its four serotypes (DENV-1 to DENV-4) infect 390 million people and are implicated in at least 25,000 deaths annually, with the largest disease burden in tropical and subtropical regions. We investigated the spatial dynamics of DENV-1, DENV-2 and DENV-3 in Brazil by applying a statistical framework to complete genome sequences. For all three serotypes, we estimated that the introduction of new lineages occurred within 7 to 10-year intervals. New lineages were most likely to be imported from the Caribbean region to the North and Northeast regions of Brazil, and then to disperse at a rate of approximately 0.5 km/day. Joint statistical analysis of evolutionary, epidemiological and ecological data indicates that aerial transportation of humans and/or vector mosquitoes, rather than Aedes aegypti infestation rates or geographical distances, determine dengue virus spread in Brazil.     

Dengue viruses activity in Piauí, Brazil

The present paper reports a laboratory investigation performed between the years of 2000 and 2002 to study a virological surveillance program introduced in the state of Piauí to support an epidemiological survey of the disease. Dengue virus type 3 (DENV-3) existence in the state was detected in May 2002 when a high number of dengue cases due to DENV-1 and DENV-2 were reported. An assessment on the population knowledge about the disease and its transmission showed that almost 50% of the population were still unaware of the epidemiological features of dengue.     

Plasma Levels of Inter-α Inhibitor Proteins in Children with Acute Dengue Virus Infection

Background

Inter-α inhibitor proteins (IaIp) belong to a family of protease inhibitors that are involved in the haemostatic and the vascular system. Dengue viruses (DENV) infections are characterized by coagulopathy and increased vascular permeability. In this study we measured the concentration of IaIp during DENV infections and evaluated its potential as a biomarker.

Methods and Findings

Concentrations of IaIp were measured in patients with acute DENV infections using a quantitative, competitive enzyme linked immunoassay. Concentrations of IaIp measured in pediatric patients suffering from severe DENV infections were significantly lower than in healthy controls.

Conclusions

This is the first report to demonstrate changes in concentration of IaIp during viral infections. The data also highlight the potential of IaIp as a biological marker for severity of DENV infections.     

Structural Changes in Dengue Virus When Exposed to a Temperature of 37°C

Previous binding studies of antibodies that recognized a partially or fully hidden epitope suggest that insect cell-derived dengue virus undergoes structural changes at an elevated temperature. This was confirmed by our cryo-electron microscopy images of dengue virus incubated at 37°C, where viruses change their surface from smooth to rough. Here we present the cryo-electron microscopy structures of dengue virus at 37°C. Image analysis showed four classes of particles. The three-dimensional (3D) map of one of these classes, representing half of the imaged virus population, shows that the E protein shell has expanded and there is a hole at the 3-fold vertices. Fitting E protein structures into the map suggests that all of the interdimeric and some intradimeric E protein interactions are weakened. The accessibility of some previously found cryptic epitopes on this class of particles is discussed.     

Investigation of the First Case of Dengue Virus Infection Acquired in Western Australia in Seven Decades: Evidence of Importation of Infected Mosquitoes?

In October 2013, a locally-acquired case of dengue virus (DENV) infection was reported in Western Australia (WA) where local dengue transmission has not occurred for over 70 years. Laboratory testing confirmed recent DENV infection and the case demonstrated a clinically compatible illness. The infection was most likely acquired in the Pilbara region in the northwest of WA. Follow up investigations did not detect any other locally-acquired dengue cases or any known dengue vector species in the local region, despite intensive adult and larval mosquito surveillance, both immediately after the case was notified in October 2013 and after the start of the wet season in January 2014. The mechanism of infection with DENV in this case cannot be confirmed. However, it most likely followed a bite from a single infected mosquito vector that was transiently introduced into the Pilbara region but failed to establish a local breeding population. This case highlights the public health importance of maintaining surveillance efforts to ensure that any incursions of dengue vectors into WA are promptly identified and do not become established, particularly given the large numbers of viraemic dengue fever cases imported into WA by travellers returning from dengue-endemic regions.     

Immature Dengue Virus: A Veiled Pathogen?

Cells infected with dengue virus release a high proportion of immature prM-containing virions. In accordance, substantial levels of prM antibodies are found in sera of infected humans. Furthermore, it has been recently described that the rates of prM antibody responses are significantly higher in patients with secondary infection compared to those with primary infection. This suggests that immature dengue virus may play a role in disease pathogenesis. Interestingly, however, numerous functional studies have revealed that immature particles lack the ability to infect cells. In this report, we show that fully immature dengue particles become highly infectious upon interaction with prM antibodies. We demonstrate that prM antibodies facilitate efficient binding and cell entry of immature particles into Fc-receptor-expressing cells. In addition, enzymatic activity of furin is critical to render the internalized immature virus infectious. Together, these data suggest that during a secondary infection or primary infection of infants born to dengue-immune mothers, immature particles have the potential to be highly infectious and hence may contribute to the development of severe disease.     

Spatio-Temporal Tracking and Phylodynamics of an Urban Dengue 3 Outbreak in S?o Paulo,Brazil

The dengue virus has a single-stranded positive-sense RNA genome of ∼10.700 nucleotides with a single open reading frame that encodes three structural (C, prM, and E) and seven nonstructural (NS1, NS2A, NS2B, NS3, NS4A, NS4B, and NS5) proteins. It possesses four antigenically distinct serotypes (DENV 1–4). Many phylogenetic studies address particularities of the different serotypes using convenience samples that are not conducive to a spatio-temporal analysis in a single urban setting. We describe the pattern of spread of distinct lineages of DENV-3 circulating in São José do Rio Preto, Brazil, during 2006. Blood samples from patients presenting dengue-like symptoms were collected for DENV testing. We performed M-N-PCR using primers based on NS5 for virus detection and identification. The fragments were purified from PCR mixtures and sequenced. The positive dengue cases were geo-coded. To type the sequenced samples, 52 reference sequences were aligned. The dataset generated was used for iterative phylogenetic reconstruction with the maximum likelihood criterion. The best demographic model, the rate of growth, rate of evolutionary change, and Time to Most Recent Common Ancestor (TMRCA) were estimated. The basic reproductive rate during the epidemics was estimated. We obtained sequences from 82 patients among 174 blood samples. We were able to geo-code 46 sequences. The alignment generated a 399-nucleotide-long dataset with 134 taxa. The phylogenetic analysis indicated that all samples were of DENV-3 and related to strains circulating on the isle of Martinique in 2000–2001. Sixty DENV-3 from São José do Rio Preto formed a monophyletic group (lineage 1), closely related to the remaining 22 isolates (lineage 2). We assumed that these lineages appeared before 2006 in different occasions. By transforming the inferred exponential growth rates into the basic reproductive rate, we obtained values for lineage 1 of R₀ = 1.53 and values for lineage 2 of R₀ = 1.13. Under the exponential model, TMRCA of lineage 1 dated 1 year and lineage 2 dated 3.4 years before the last sampling. The possibility of inferring the spatio-temporal dynamics from genetic data has been generally little explored, and it may shed light on DENV circulation. The use of both geographic and temporally structured phylogenetic data provided a detailed view on the spread of at least two dengue viral strains in a populated urban area.     

Interplay of RNA Elements in the Dengue Virus 5′ and 3′ Ends Required for Viral RNA Replication

Dengue virus (DENV) is a member of the Flavivirus genus of positive-sense RNA viruses. DENV RNA replication requires cyclization of the viral genome mediated by two pairs of complementary sequences in the 5′ and 3′ ends, designated 5′ and 3′ cyclization sequences (5′-3′ CS) and the 5′ and 3′ upstream of AUG region (5′-3′ UAR). Here, we demonstrate that another stretch of six nucleotides in the 5′ end is involved in DENV replication and possibly genome cyclization. This new sequence is located downstream of the AUG, designated the 5′ downstream AUG region (5′ DAR); the motif predicted to be complementary in the 3′ end is termed the 3′ DAR. In addition to the UAR, CS and DAR motifs, two other RNA elements are located at the 5′ end of the viral RNA: the 5′ stem-loop A (5′ SLA) interacts with the viral RNA-dependent RNA polymerase and promotes RNA synthesis, and a stem-loop in the coding region named cHP is involved in translation start site selection as well as RNA replication. We analyzed the interplay of these 5′ RNA elements in relation to RNA replication, and our data indicate that two separate functional units are formed; one consists of the SLA, and the other includes the UAR, DAR, cHP, and CS elements. The SLA must be located at the 5′ end of the genome, whereas the position of the second unit is more flexible. We also show that the UAR, DAR, cHP, and CS must act in concert and therefore likely function together to form the tertiary RNA structure of the circularized DENV genome.Dengue virus (DENV), a member of the Flaviviridae family, is a human pathogen causing dengue fever, the most common mosquito-borne viral disease in humans. The virus has become a major international public health concern, with 3 billion people at risk for infection and an estimated 50 million dengue cases worldwide every year (28). Neither specific antiviral therapies nor licensed vaccines are available, and the biology of the virus is poorly understood.DENV is a small enveloped virus containing a positive-stranded RNA genome with a length of approximately 10.7 kb. The virus encodes one large polyprotein that is co- and posttranslationally cleaved into 10 viral proteins. The structural proteins C, prM/M, and E are located in the N terminus, followed by the nonstructural proteins NS1, NS2A, NS2B, NS3, NS4A, NS4B, and NS5 (6, 10). NS5, the largest of the viral proteins, functions as an RNA-dependent RNA polymerase (RdRP) (29). The coding region is flanked at both ends by untranslated regions (UTR). The 5′ end has a type I cap structure (m⁷GpppAmp) mediating cap-dependent translation, but the virus can switch to a noncanonical translation mechanism under conditions in which translation factors are limiting (13). Cellular mRNAs are known to circularize via a protein-protein bridge between eIF4G and eIF4E (the cap binding complex) at the 5′ end and the poly(A) binding protein (PABP) at the 3′ end, enhancing translation efficiency. Despite the fact that the DENV 3′ UTR lacks a poly(A) tail, recent findings demonstrated binding of PABP to the 3′ UTR and an effect on RNA translation, suggesting a similar mechanism (12, 26).In addition to a presumed protein-mediated genome circularization regulating viral translation, an RNA-RNA-based 5′ and 3′ (5′-3′) end interaction, which can occur in the absence of proteins, leads to circularization of the viral genome (1, 3, 4, 18, 20, 30, 33, 34). This cyclization of the genome is necessary for viral RNA replication, and thus far, two complementary sequences at the 5′ and 3′ ends have been identified (3). The first are the cyclization sequences (CS) present in the capsid-coding region at the 5′ end (5′ CS) and upstream of the 3′ stem-loop (3′ SL) in the 3′ UTR (3′ CS) (2, 4, 18, 20, 30). A second sequence, known as the 5′ upstream AUG region (5′ UAR) element in the 5′ UTR, base pairs with its complementary 3′ UAR counterpart, which is located at the bottom part of 3′ SL (1, 4, 30). Recently, the structure of the 5′ end of the DENV genome hybridized to the 3′ end was determined in solution (25), confirming previous computer-predicted structures for genome cyclization (4, 20, 30). Besides the base pairing between 5′-3′ UAR and 5′-3′ CS sequences, a third stretch of nucleotides was identified to form a double-stranded (ds) region between the 5′ and 3′ ends.In addition to RNA sequences involved in 5′-3′-end interactions that are necessary for cyclization, the 5′ end of the viral genome harbors at least two more functional RNA elements, the stem-loop A (SLA) and capsid-coding region hairpin (cHP). The SLA consists of the first 70 nucleotides (nt) of the genome, forming a stable stem-loop structure. This structure has been confirmed in several studies and identified as a promoter element for RNA synthesis that recruits the viral RdRp NS5 (16, 22). Once NS5 is bound to the SLA at the 5′ end, it is believed to be delivered to the initiation site of minus-strand RNA synthesis at the 3′ end via 5′-3′ RNA-RNA circularization. In addition, a short poly(U) tract located immediately downstream of SLA has been shown to be necessary for RNA synthesis, although it is not involved in genome circularization (22). Finally, the cHP element resides within the capsid-coding region; it directs start codon selection and is essential for RNA replication (8, 9). The cHP structure is more important than its primary sequence. For start codon selection, it is believed that the cHP stalls the scanning initiation complex over the first AUG, favoring its recognition (9). In the case of RNA replication, the cHP likely stabilizes the overall 5′-3′ panhandle structure or participates in recruitment of factors associated with the replicase machinery (8).In this study, we demonstrate that in addition to the 5′ CS and 5′ UAR sequences, a third stretch of nucleotides in the 5′ end is required for RNA replication and appears to be involved in genome circularization. This new motif is located downstream of the AUG and was therefore designated the downstream AUG region (5′ DAR) element, with the predicted counterpart in the 3′ end designated the 3′ DAR. Our results indicate that the 5′ DAR modulates RNA-RNA interaction and RNA replication, and restoring complementarity between the 5′ DAR and 3′ DAR rescues detrimental effects caused by mutations in the 5′ DAR on genome circularization and RNA replication. Although the role of the predicted 3′ DAR counterpart is less conclusive, it may serve to make other structures and sequences in the 3′ end available for 5′-3′ RNA-RNA interaction to facilitate the replication-competent conformation of the DENV genome.Furthermore, we analyzed the functional interplay of RNA elements in the viral 5′ end, showing that two separate units are formed during replication. The first consists of the SLA, and it must be located at the very 5′ end of the genome. The second unit includes UAR, DAR, cHP, and CS elements, and the positional requirements are more flexible within the DENV RNA 5′ terminus. However, all four elements in the second unit must act in concert, forming a functional tertiary RNA structure of the circularized viral genome.     

Spatial Distribution of the Risk of Dengue and the Entomological Indicators in Sumaré, State of S?o Paulo,Brazil

Dengue fever is a major public health problem worldwide, caused by any of four virus (DENV-1, DENV-2, DENV-3 and DENV-4; Flaviviridae: Flavivirus), transmitted by Aedes aegypti mosquito. Reducing the levels of infestation by A. aegypti is one of the few current strategies to control dengue fever. Entomological indicators are used by dengue national control program to measure the infestation of A. aegypti, but little is known about predictive power of these indicators to measure dengue risk. In this spatial case-control study, we analyzed the spatial distribution of the risk of dengue and the influence of entomological indicators of A. aegypti in its egg, larva-pupa and adult stages occurring in a mid-size city in the state of São Paulo. The dengue cases were those confirmed by the city''s epidemiological surveillance system and the controls were obtained through random selection of points within the perimeter of the inhabited area. The values of the entomological indicators were extrapolated for the entire study area through the geostatistical ordinary kriging technique. For each case and control, the respective indicator values were obtained, according with its geographical coordinates and analyzed by using a generalized additive model. Dengue incidence demonstrated a seasonal behavior, as well as the entomological indicators of all mosquito''s evolutionary stages. The infestation did not present a significant variation in intensity and was not a limiting or determining factor of the occurrence of cases in the municipality. The risk maps of the disease from crude and adjusted generalized additive models did not present differences, suggesting that areas with the highest values of entomological indicators were not associated with the incidence of dengue. The inclusion of other variables in the generalized additive models may reveal the modulatory effect for the risk of the disease, which is not found in this study.     

Live Cell Analysis and Mathematical Modeling Identify Determinants of Attenuation of Dengue Virus 2’-O-Methylation Mutant

Dengue virus (DENV) is the most common mosquito-transmitted virus infecting ~390 million people worldwide. In spite of this high medical relevance, neither a vaccine nor antiviral therapy is currently available. DENV elicits a strong interferon (IFN) response in infected cells, but at the same time actively counteracts IFN production and signaling. Although the kinetics of activation of this innate antiviral defense and the timing of viral counteraction critically determine the magnitude of infection and thus disease, quantitative and kinetic analyses are lacking and it remains poorly understood how DENV spreads in IFN-competent cell systems. To dissect the dynamics of replication versus antiviral defense at the single cell level, we generated a fully viable reporter DENV and host cells with authentic reporters for IFN-stimulated antiviral genes. We find that IFN controls DENV infection in a kinetically determined manner that at the single cell level is highly heterogeneous and stochastic. Even at high-dose, IFN does not fully protect all cells in the culture and, therefore, viral spread occurs even in the face of antiviral protection of naïve cells by IFN. By contrast, a vaccine candidate DENV mutant, which lacks 2’-O-methylation of viral RNA is profoundly attenuated in IFN-competent cells. Through mathematical modeling of time-resolved data and validation experiments we show that the primary determinant for attenuation is the accelerated kinetics of IFN production. This rapid induction triggered by mutant DENV precedes establishment of IFN-resistance in infected cells, thus causing a massive reduction of virus production rate. In contrast, accelerated protection of naïve cells by paracrine IFN action has negligible impact. In conclusion, these results show that attenuation of the 2’-O-methylation DENV mutant is primarily determined by kinetics of autocrine IFN action on infected cells.     

Epidemiological Trends of Dengue Disease in Brazil (2000–2010): A Systematic Literature Search and Analysis

A literature survey and analysis was conducted to describe the epidemiology of dengue disease in Brazil reported between 2000 and 2010. The protocol was registered on PROSPERO (CRD42011001826: http://www.crd.york.ac.uk/prospero/display_record.asp?ID=CRD42011001826). Between 31 July and 4 August 2011, the published literature was searched for epidemiological studies of dengue disease, using specific search strategies for each electronic database. A total of 714 relevant citations were identified, 51 of which fulfilled the inclusion criteria. The epidemiology of dengue disease in Brazil, in this period, was characterized by increases in the geographical spread and incidence of reported cases. The overall increase in dengue disease was accompanied by a rise in the proportion of severe cases. The epidemiological pattern of dengue disease in Brazil is complex and the changes observed during this review period are likely to have been influenced by multiple factors. Several gaps in epidemiological knowledge regarding dengue disease in Brazil were identified that provide avenues for future research, in particular, studies of regional differences, genotype evolution, and age-stratified seroprevalence.

Systematic Review Registration

PROSPERO registration number: CRD42011001826.     

Crystal Structure of the Dengue Virus Methyltransferase Bound to a 5′-Capped Octameric RNA

The N-terminal domain of the flavivirus NS5 protein functions as a methyltransferase (MTase). It sequentially methylates the N7 and 2′-O positions of the viral RNA cap structure (GpppA→^7meGpppA→^7meGpppA_2′-O-me). The same NS5 domain could also have a guanylyltransferase activity (GTP+ppA-RNA→GpppA). The mechanism by which this protein domain catalyzes these three distinct functions is currently unknown. Here we report the crystallographic structure of DENV-3 MTase in complex with a 5′-capped RNA octamer (G_pppAGAACCUG) at a resolution of 2.9 Å. Two RNA octamers arranged as kissing loops are encircled by four MTase monomers around a 2-fold non-crystallography symmetry axis. Only two of the four monomers make direct contact with the 5′ end of RNA. The RNA structure is stabilised by the formation of several intra and intermolecular base stacking and non-canonical base pairs. The structure may represent the product of guanylylation of the viral genome prior to the subsequent methylation events that require repositioning of the RNA substrate to reach to the methyl-donor sites. The crystal structure provides a structural explanation for the observed trans-complementation of MTases with different methylation defects.     

Re-Emergence of Dengue Virus Type 3 in Canton,China, 2009–2010, Associated with Multiple Introductions through Different Geographical Routes

Background

Endemic dengue virus type 3 (DENV-3) infections have not been reported in Canton, China, since 1980. In March 2009, DENV-3 was isolated for the second time, occurring about 30 years after the previous circulation. In August, 3 other cases emerged. One much larger outbreak occurred again in 2010. To address the origin and particularly to determine whether the outbreaks were caused by the same viral genotype, we investigated the epidemiological and molecular characteristics of the introduction, spread and genetic microevolution of DENV-3 involved.

Methodology/Principal Findings

Three imported cases (index-1,2,3) separately traveled back from Vietnam, India and Tanzania, resulted in 1, 3 and 60 secondary autochthonous cases, respectively. In autochthonous cases, 64.6% positive in IgM anti-DENV and 18.6% in IgG from a total of 48 submitted serum samples, accompanied by 7 DENV-3 isolates. With 99.8%, 99.7%, and 100% envelope gene nucleotidic identity, 09/GZ/1081 from index-1 and endemic strain (09/GZ/1483) belonged to genotype V; 09/GZ/10616 from index-2 and endemic strains (09/GZ/11144 and 09/GZ/11194) belonged to genotype III Clade-A; and 10/GZ/4898 from index-3 and all four 2010 endemic DENV-3 strains belonged to genotype III Clade-B, respectively.

Conclusions/Significance

Both epidemiological and phylogenetic analyses showed that the 2010 outbreak of dengue was not a reemergence of the 2009 strain. Introductions of different genotypes following more than one route were important contributory factors for the 2009–2010 dengue epidemics/outbreaks in Canton. These findings underscore the importance of early detection and case management of imported case in preventing large-scale dengue epidemics among indigenous peoples of Canton.     

Is the Gut the Major Source of Virus in Early Simian Immunodeficiency Virus Infection?

The acute phases of human immunodeficiency virus (HIV) and simian immunodeficiency virus (SIV) infection are characterized by rapid and profound depletion of CD4⁺ T cells from the guts of infected individuals. The large number of CD4⁺ T cells in the gut (a large fraction of which are activated and express the HIV/SIV coreceptor CCR5), the high level of infection of these cells, and the temporal coincidence of this CD4⁺ T-cell depletion with the peak of virus in plasma in acute infection suggest that the intestinal mucosa may be the major source of virus driving the peak viral load. Here, we used data on CD4⁺ T-cell proportions in the lamina propria of the rectums of SIV-infected rhesus macaques (which progress to AIDS) and sooty mangabeys (which do not progress) to show that in both species, the depletion of CD4⁺ T cells from this mucosal site and its maximum loss rate are often observed several days before the peak in viral load, with few CD4⁺ T cells remaining in the rectum by the time of peak viral load. In contrast, the maximum loss rate of CD4⁺ T cells from bronchoalveolar lavage specimens and lymph nodes coincides with the peak in virus. Analysis of the kinetics of depletion suggests that, in both rhesus macaques and sooty mangabeys, CD4⁺ T cells in the intestinal mucosa are a highly susceptible population for infection but not a major source of plasma virus in acute SIV infection.The acute phase of human immunodeficiency virus (HIV) infection is characterized by moderate CD4⁺ T-cell depletion in blood, followed by a transient partial restoration of CD4⁺ T-cell numbers and eventually by a slow long-term CD4⁺ T-cell decline in the chronic phase that lasts for several years. Studies of CD4⁺ T-cell depletion in mucosal sites, often conducted with simian immunodeficiency virus (SIV)-infected macaques, have demonstrated that mucosal CD4⁺ T-cell depletion is more rapid and profound (3, 10, 13, 19, 21). The severe depletion of cells in the gut in early infection is thought to be driven in part by the phenotype of the cells present, which are predominantly CCR5⁺ and in general more activated than their circulating counterparts. As such, these mucosal CD4⁺ T cells are highly susceptible to productive infection with the dominant CCR5-tropic strains of HIV and SIV present in early infection (20). The rapid depletion of CD4⁺ T cells at mucosal sites is accompanied by relatively high numbers of infected cells (10, 13) and is temporally associated with the peak viral load in plasma, suggesting that the infection of mucosal CD4⁺ T cells may be responsible for the majority of virus replication occurring during acute infection (10, 15, 21, 22).The size of the CD4⁺ T-cell pool in the gut is a matter of some controversy, with estimates ranging from ∼5 to 50% of the total body pool of these cells (reviewed in reference 5). Regardless of the precise numbers, the gut (and particularly the mucosal lamina propria) contains a significant proportion of the body CD4⁺ CCR5⁺ memory T cells, which are depleted very early in infection. However, whether CD4⁺ T cells in the gut are merely a target of early infection or whether they are a major driver of early viral growth and peak viral loads in acute infection is unclear. Here we use a combination of experimental data and modeling to demonstrate that the gut is unlikely to be a major source of virus production in acute SIV infection.     

Can Human Movements Explain Heterogeneous Propagation of Dengue Fever in Cambodia?

Background

Determining the factors underlying the long-range spatial spread of infectious diseases is a key issue regarding their control. Dengue is the most important arboviral disease worldwide and a major public health problem in tropical areas. However the determinants shaping its dynamics at a national scale remain poorly understood. Here we describe the spatial-temporal pattern of propagation of annual epidemics in Cambodia and discuss the role that human movements play in the observed pattern.

Methods and Findings

We used wavelet phase analysis to analyse time-series data of 105,598 hospitalized cases reported between 2002 and 2008 in the 135 (/180) most populous districts in Cambodia. We reveal spatial heterogeneity in the propagation of the annual epidemic. Each year, epidemics are highly synchronous over a large geographic area along the busiest national road of the country whereas travelling waves emanate from a few rural areas and move slowly along the Mekong River at a speed of ∼11 km per week (95% confidence interval 3–18 km per week) towards the capital, Phnom Penh.

Conclusions

We suggest human movements – using roads as a surrogate – play a major role in the spread of dengue fever at a national scale. These findings constitute a new starting point in the understanding of the processes driving dengue spread.