From January to December 1997, 40,237 serum samples were drawn from febrile patients in Belém, 20,038 (49.8%) of whom were male. The patients were all residents of the municipalities of Belém (31,506 samples) and Ananindeua (8,731 samples). Most were ambulatory patients seen in the Arbovirus Department, Evandro Chagas Institute; some patients were referred by public health centers, private physicians, and clinics. At the Institute, all patients were examined clinically and had blood samples drawn. A questionnaire was administered that included information about clinical symptoms and signs and demographic data. Patients who had been ill <10 days and whose serologic tests were negative for dengue were requested to provide a second blood sample 7 to 14 days later. For patients who had at least one hemorrhage and either dehydration or hemoconcentration, a leukogram, platelet count, and hematocrit were performed.

Serum samples were tested for dengue antibodies by MAC ELISA (1) and hemagglutination-inhibition test (7). The antigens for both tests were prepared by using infected mouse brain extracted by the sucrose acetone method. The criteria used for establishing primary and secondary infection were those recommended by the World Health Organization (7,8).

To isolate dengue virus, 0.1-ml aliquots of whole blood from patients with clinical symptoms lasting <5 days and negative serologic results were injected into cultures of C6/36 cells (10). The cultures were visually examined daily, and cells were tested on days 7 and 14 by immunofluorescence (3).

Isolated virus strains were initially screened by direct immunofluorescence against a flavivirus standard hyperimmune fluid prepared at the Institute. Strains that reacted were identified to serotype by using an indirect immunofluorescence test with monoclonal antibodies against the four dengue viruses provided by the Centers for Disease Control and Prevention.

The epidemic distribution was accompanied early in 1997 by a seasonal increase in rainfall typical of the Brazilian Amazon. However, the highest dengue positivity rates were reported in the dry months of September to December. Coincidentally, DEN-2 virus was isolated in October. At first, cases were reported only in Belém and Ananindeua, but in December 1997, at least 15 other municipalities reported transmission cycles involving either DEN-1 or DEN-2 or both.

Figure

Figure. Monthly distribution of rain precipitation (line) and Dengue fever cases as seen at Instituto Evandro Chagas, Pará, Brazil. Source for rain precipitation, Instituto Nacional de Meteorologia, Brazil.

Of all sera collected, 17,525 (43.5%) were positive by serologic testing or DEN virus isolation: 13,805 (78.8%) from Belém and 3,720 (21.2%) cases from Ananindeua (Figure). Both primary and secondary serologic responses were found. Among the positive samples, 9,469 (54.25%) were among female and 7,971 (45.75%) among male patients (p < 0.0001); 45.9% of the patients were 25 to 44 years of age (Table 1). Paired serum samples were obtained from 3,558 patients, 2,997 (84.2%) of whom had serologic conversions.

No DHF cases were reported, in spite of the fact that since October 1997 DEN-2 had been isolated from several 24 patients with previous DEN-1 infection (Table 2). Clinical symptoms were more severe in older children and adults than in young children, but no differences were observed between the clinical symptoms of patients with DEN-1 and DEN-2 infection.

During 1953 and 1954, Causey and Theiler (11) used seroneutralization in mice to survey several municipalities of the Amazon region; they found 9.8% and 2.2% of serum samples positive for dengue virus serotypes DEN-1 and DEN-2, respectively, among residents 50 years of age. These results suggest that dengue viruses were circulating in Belém and other municipalities of the Amazon Valley early in the 20th century.

The dengue epidemic in Belém was unusual in that a "lag phase" (12) lasted for at least 4 months before extensive transmission began. In spite of laboratory diagnosis of cases from the beginning of the outbreak and the reporting of these results to health authorities, control measures were unsuccessful. Consequently, explosive transmission began in March 1997 and is still occurring (1998-99).

In the last 4 years, the Brazilian Ministry of Health (13) has reported an increase in dengue cases from approximately 56,000 in 1994 to >530,000 in 1998. High indexes of Ae. aegypti infestation exist in all important urban centers of Brazil, reflecting a poor health education program. The fact that more cases occurred in female than in male patients (p <0.0001) suggests that women are at increased risk for dengue infection because of high peridomestic exposure.

Dengue transmission has been reported in 22 of the 27 Brazilian states, and the mosquito vector is present in all states; therefore, the situation in Brazil may be rapidly approaching hyperendemicity, with the cocirculation of two serotypes (DEN-1 since 1986 and DEN-2 since 1990). The risk for DHF will increase if a new serotype (DEN-3 or DEN-4) is introduced. Endemicity is the most constant factor associated with the evolution of epidemic DHF in a geographic area (14). Although few cases have been reported until recently, DHF may become an important cause of hospitalization and death in the Americas, including Brazil.