At HCB, as part of the routine workup for febrile patients, serum samples were collected, stored at −20°C, and transported in liquid nitrogen to Quito for PCR. Hospital personnel also tested these samples by using a dengue IgM/IgG-capture ELISA with high specificity and sensitivity (Panbio, Waltham, MA, USA). At the SNEM laboratory, 2–4 drops of whole blood were collected on filter paper (Whatman 903, Kent, UK) from persons spontaneously seeking or referred for malaria diagnosis. Filter papers were left to dry overnight, stored at 0°C in a zipper bag, and transported to Quito for analysis. DENV was amplified from total RNA from blood spots and serum as described (12,13). Enriched media from DENV 1–4 cultures were used as positive controls. Quality of RNA extracts was tested by using β-actin gene amplification. PCR products were sized in 1.5% agarose gel electrophoresis and SYBR Safe DNA Gel Stain (Invitrogen, Carlsbad, CA, USA) (1:10,000) under UV light. PCR product was sent to Functional Biosciences (Madison, WI, USA) for sequencing. All participants provided oral consent. The study was approved by institutional review board committees at the Universidad San Francisco de Quito (Quito, Ecuador) and the University of Michigan (Ann Arbor, MI, USA).

From July 2010 through February 2011, a total of 77 samples (36 serum, 41 blood spot) were collected from febrile patients. Of these 36 hospital patients, 10 were from Borbón and 26 were from the surrounding communities; patients were 2–74 years of age (median 19 years), and fever duration was 3–20 days (median 6 days). Six (17%) of the 36 serum samples and 7 (17%) of the 41 blood spots were positive for dengue by PCR; whereas 10 (29%) of the 34 serum samples tested by ELISA were positive for dengue IgM (Table). The ages of the 10 patients with positive results by ELISA were 3–63 years (median 38 years), and fever duration was 4–13 days (median 6.5 days). Although positive PCRs are sufficient indicators of acute infection, the positive IgM reflects a cumulative incidence over 60–90 days. Therefore, the 17% PCR-positive estimate is a lower bound, and the 42% estimate by PCR or ELISA is an upper bound. The DENVs in the 13 samples that were positive by PCR were either serotype 2 or 3; samples from 1 patient contained both serotypes, probably resulting from a co-infection. Each institution (HCB and SNEM) submitted samples from the same 2 patients, but only 1 was positive for dengue by PCR. During this same period, DENV-1, -2, and -4 were detected in coastal cities in other regions of Ecuador (DENV-3 has not been reported in Ecuador since 2009 [14]). These data suggest that DENV-3 was introduced into the study region from another country. Colombia is a likely source, given the proximity of the study region to Colombia and the fact that DENV-3 was circulating in Colombia; but without sequence data, the geographic location of the source cannot be confirmed.

This study provides evidence of a transition of febrile disease etiology from Plasmodium spp., traditionally highly prevalent in this area, to reemerging pathogens, such as DENV. Other possible fever-causing agents for which we did not test include Leptospira spp., Rickettsia spp., Brucella spp., and viral agents of encephalitis (11,15). Given that malaria has historically been the predominant cause of fever in the developing world, febrile patients are often triaged toward malaria treatment, especially in resource-poor areas. Moreover, the awareness of malaria is embedded culturally and behaviorally in these communities: having a fever is equated with having malaria. Of the 40 febrile patients who sought treatment for malaria, none had malaria but 17% had dengue (positive for DENV by PCR). Further research is needed to refine this estimate and explore how the Ministry of Health and the general population should be directed to manage cases of fever during this etiologic transition. The need to bring attention to this transition is exemplified by our finding of DENV-3 in our study site; this serotype has not been isolated in other parts of Ecuador since 2009. Esmeraldas, therefore, may be a major source of newly introduced dengue serotypes into Ecuador.