Mosquitoborne Infections after Hurricane Jeanne, Haiti, 2004

After Hurricane Jeanne in September 2004, surveillance for mosquitoborne diseases in Gonaïves, Haiti, identified 3 patients with malaria, 2 with acute dengue infections, and 2 with acute West Nile virus infections among 116 febrile patients. These are the first reported human West Nile virus infections on the island of Hispaniola.


The Study
We established laboratory-based fever surveillance at the 3 clinics providing healthcare in Gonaïves after the passage of Hurricane Jeanne. Febrile patients (core temperature ≥38.5°C when first assessed) were asked to provide blood for a serum sample and thick and thin malaria smears. The attending physician recorded each patient's medical history, conducted a physical examination, and reported the discharge diagnosis and the therapy that was provided. We asked patients to return in 2 weeks so that a convalescent-phase serum sample could be collected.
Malaria smears were stained and read by using standard methods (1) at CDC. To diagnose dengue, we used nested PCR and the TaqMan assay to detect dengue viral RNA in serum samples obtained <5 days after onset of symptoms (2,3). In addition, we used an immunoglobulin M (IgM) antibody-capture (MAC)-ELISA to detect antidengue IgM antibodies in all serum specimens (4) at CDC. A result was considered positive when optical density, after comparison to negative serum and control antigen, was >0.20. All serum specimens were also tested for the presence of IgG antibodies to determine previous exposure to flaviviruses by using an IgG ELISA. In paired samples, a full titration of 4-fold dilutions of serum was used. The endpoint titration of IgG was determined to assess seroconversion (5). Each plate was compared with a negative control serum specimen. Because of cross-reactivity between anti-flavivirus antibodies, we used a microspherebased immunoassay (MIA) with a quadratic discrimination analysis (6) and a plaque reduction neutralization test (PRNT) to distinguish between infecting flaviviruses. For the PRNT, serial dilutions of heat-inactivated serum were incubated with defined amounts of West Nile, Saint Louis encephalitis, and dengue viruses 1-4 for 2 hours at room temperature. The nonneutralized viral fraction was subsequently adsorbed onto a monolayer of Vero cells for 1 hour. The resultant plaques were counted and compared with results for the control virus with no serum. The endpoint of the titration was the highest dilution of serum that reduced the number of plaques 90% compared with the control results.
Of the 116 thick and thin smears, 3 (3%) samples showed a high level of parasitemia with Plasmodium falciparum. The 3 corresponding patients had fever with no apparent source. Malaria was suspected in 2 of the patients by their clinical symptoms; the third patient was thought to have typhoid and was treated with trimethoprimsulfamethoxazole.
Two patients (2%) had acute, secondary dengue infections that were confirmed as positive by both IgM and IgG serologic tests. Both patients had a chief report of fever with no source, but malaria was suspected by the attending physician, and 1 patient was treated with chloroquine. We were not able identify dengue viral particles in the serum specimens of these patients. However, 79 patients (68%) were positive for anti-dengue IgG, which suggests a high level of flavivirus transmission in this area in the recent past (Figure).
Two patients (2%) had MIA results consistent with acute West Nile virus infection. The results were confirmed by PRNT (Table). Both patients were febrile in the clinic; 1 was a 13-year-old boy and the other was an infant girl <1 year of age. In addition to fever, the 13-year-old reported headache and abdominal pain, while cough was reported in the infant. Acute malaria was clinically diagnosed in both patients. The older child received chloroquine, while the younger child received only acetaminophen for fever control.

Conclusions
This surveillance program was established to assess the incidence of vectorborne diseases in the wake of Hurricane Jeanne. A total of 116 acutely febrile patients had blood drawn to determine whether a mosquitoborne disease was the etiologic agent of fever. An outbreak of mosquitoborne disease was not detected during the period of surveillance. Our data are consistent with previously published reports, which indicate that the incidence of arboviral infections rarely increases after water-related dis-asters (e.g., floods, hurricanes) (7−9). However, malaria outbreaks are common in such settings (9,10).
Despite the absence of an outbreak, our surveillance did identify the ongoing transmission of 3 mosquitoborne pathogens. Specifically, we diagnosed 3 cases of acute malaria, 2 cases of acute dengue, and 2 cases of acute West Nile virus infection. We also detected a high seroprevalence of dengue infections in children, which suggests substantial local dengue transmission in the Gonaïves area in the recent past.
The high seroprevalence of dengue and the low smearpositive rate of malaria from our surveillance were consistent with previously reported studies in this region of Haiti (11,12) The fact that the rate of West Nile virus infection was equal to the rate of acute dengue infection among our participants is of concern. Moreover, because both viruses can cause a nonlocalizing fever, the potential for confusion with malaria exists. Differentiating the cause of acute nonlocalizing febrile illnesses by examining malaria smears before initiating therapy, especially in an area with a history of low smear positivity, is therefore important.