Figure 1

Figure 1. Map of study sites in Ecuador, Peru, and Venezuela. The study in Peru and Ecuador was operated jointly by the US Naval Medical Research Center Detachment with each country’s Institutes of...

Patients with acute, undifferentiated, febrile illness were recruited into a clinic-based surveillance study conducted jointly by the US Naval Medical Research Center Detachment (NMRCD) and the Peruvian and Ecuadorian Institutes of Health (Figure 1). Study protocols (NMRCD.2000.0006 [Peru] and NMRCD.2001.0002 [Ecuador]) were approved by the Naval Medical Research Center Institutional Review Boards in compliance with all US federal regulations governing the protection of human subjects.

Patient sera were injected onto African green monkey Vero cells or Ae. albopictus C6/36 cells and examined for a range of arboviruses, including all 4 DENV serotypes, by immunofluorescent assay. From 2000 through 2008, DENV-3 (1,572 isolates) was the dominant serotype in circulation in the study sites, followed by DENV-1 (205 isolates) and DENV-2 (87 isolates). From the initiation of the study in May 2000 until February 2006, DENV-4 circulation was rarely detected in the NMRCD-affiliated study sites in either country; the exceptions were a small number of isolates in 2000 in Tumbes, Peru (n = 2) and Guyaquil, Ecuador (n = 6).

Low-level DENV-4 transmission was again detected in Ecuador and coastal Peru during 2006 and 2007, in isolates from patients in Tumbes, Trujillo, and Guayaquil, none of whom had reported recent history of travel outside their respective areas. DENV-4 continued to circulate in sites along the northern coast of Peru in 2008 and 2009, in co-circulation with DENV-1. In February 2008, DENV-4 spread to the cities of Iquitos and Yurimaguas, located in the Loreto Department in the tropical rainforest region of northeastern Peru. By October 2008, DENV-4 had nearly completely displaced DENV-3, which had been the only serotype detected in the region during the previous 3 years. Nine (56%) of 16 DENV isolates obtained from febrile patients in August 2008, 55 (85%) of 65 isolates obtained in September 2008, and 305 (98%) of 311 isolates obtained from October 2008 through February 2009 were DENV-4. After the introduction of DENV-4, the total number of DENV-3 isolates decreased during peak months of DENV transmission (October through February) from 176 during 2006–2007 and 420 during 2007–2008 to <10 isolates during the same period in 2008–2009. More recently (March 2009), DENV-4 strains have spread south to Lima, the capital city of Peru, causing a small, localized outbreak on a military base.

To characterize the DENV-4 isolates, a 1,485-bp sequence covering the entire mature envelope (E) gene was amplified and sequenced from a representative set of viruses from Guayaquil (n = 6), Tumbes (n = 6), Piura (n = 2), Trujillo (n = 1), Iquitos (n = 9), Yurimaguas (n = 7), and Lima (n = 2), all collected during 2000 through 2009. Isolates collected in Peru since 2006 shared >99.5% nucleotide identity but exhibited <97% identity with DENV-4 strains collected in Ecuador and coastal Peru in 2000. At the amino acid level, the 2006–2009 isolates were nearly invariant, with <1 amino acid difference.

Figure 2

Figure 2. Phylogenetic analysis of the envelope gene of dengue virus serotype 4 (DENV–4) strains from Ecuador, Peru, and Venezuela. Similar topologies were observed from neighbor-joining (depicted), maximum likelihood, and maximum parsimony analyses,...

For further characterization, the DENV-4 strains from Ecuador and Peru were compared with DENV-4 sequences from Latin America (9,10) and Southeast Asia (11) available from the GenBank database. In addition, to provide a wider array of recent isolates from the Caribbean region, DENV-4 isolated from febrile patients in Aragua State, Venezuela (n = 15) collected during 2000 through 2007 were sequenced. Based on phylogenetic analyses, all DENV-4 isolates belong to genotype II (data not shown), although the 2006–2009 isolates segregated into a markedly different clade than the strains from 2000 (Figure 2). The 2000 isolates clustered more closely with a previous 1994 Ecuador isolate (9), related to the initial 1981 Caribbean introduction DENV-4 strains (designated as subtype A). The 2006–2009 isolates were most closely related to recent DENV-4 isolates from Venezuela, with as low as 0.8% nucleotide divergence, and formed a lineage distinct from previously published DENV-4 Caribbean basin strains (9,10), with strong bootstrap support (Figure 2). This lineage is distinguished from previously reported DENV-4 genotype II strains by 3 conserved amino acid variations in the E protein: S64L, A235T, and S403A.

We report DENV-4 expanding rapidly through northern Peru, particularly in the Loreto Department in the tropical rainforest region, spreading through a population immunologically naïve to this serotype. In the past 2 decades, populations in northern Peru have been exposed to the other 3 DENV serotypes, thus increasing the possibility for severe disease, including DHF. In Iquitos, DHF was first reported during a DENV-3 epidemic in 2004 (M. Sihuincha and C. Rocha, pers. comm.). DHF had not been observed after the introduction of DENV-2 despite large numbers of infected residents (13), presumably due to cross-protection afforded by prior infection with DENV-1 (14). For the currently circulating lineage of DENV-4, the levels of either cross-protection or antibody-dependent enhancement of infection resulting from prior heterologous infection remain to be determined.

The mechanisms responsible for DENV-3 displacement in Loreto are unclear. Following several years of DENV-3 circulation in the region, serotype-specific antibody prevalence is high (≈45% of the population, based on virus neutralization assays [T.J. Kochel, unpub. data]). However, it is unlikely that herd immunity is sufficient to explain the dramatic decrease in DENV-3 transmission. The presence of broadly cross-protective antibodies during the acute and early convalescent phases following DENV-4 infection, when combined with the large number of DENV-3 immune persons, could suppress transmission of DENV-3 strains (15). Another possibility is serotype competition within the vector species. Analysis of virus strains from Ae. aegypti mosquitoes collected during the transitional period could help elucidate whether such intertypic competition is occurring.

Genetically, the 2006–2009 isolates analyzed in this study do not appear to be related to viruses collected from Ecuador and northern Peru during 2000, which were similar to the initial Latin American introduction strains (Figure 2). Instead the 2006–2009 isolates were most closely related to viruses from Venezuela collected during 2000 through 2007, forming a lineage distinct from the DENV-4 genotype II B4 lineage (10). Genetic conservation among isolates from Peru and similarity with isolates from Venezuela support an introduction event into northern Peru and Ecuador from the northern region of South America before 2006, and a subsequent introduction from coastal Peru into Loreto, although more data from other regions of South America and the Caribbean basin would be necessary to more clearly delineate the geographic spread of this virus strain.