During 2001–2007, a total of 4 communities (3 in Los Santos Province and 1 in Veraguas Province) within hantavirus-endemic agroecosystems in western Panama were sampled 2–4 times at 1- to 3-year intervals (Table 1). Informed written consent was obtained from all adult participants and from parents or legal guardians of minors. Consent and assent forms were reviewed and approved by institutional ethics review boards at the University of New Mexico, the Gorgas Memorial Institute in Panama City, and the protocol review committee of the International Centers for Infectious Diseases Research program of the National Institute of Allergy and Infectious Diseases. Eligible participants were all adults and children >2 years of age who permanently resided in each community according to the 2000 national census. The reasons for noninclusion in the first and subsequent surveys were absence during the week of the survey and refusal to participate.

A questionnaire administered to the head of household asked for a history of respiratory-related illnesses and hospitalizations within the past 3 years. Venous blood was collected from all family members for serologic testing. Results of the surveys were provided to each participating community through community meetings. Surveillance for HPS was conducted in the same communities as the serosurvey and nationally through reports to the Ministry of Health, and cases of HPS were confirmed by questionnaire. The diagnosis of HPS required finding immunoglobulin (Ig) M in acute-phase serum, detection of Choclo virus RNA in serum by reverse transcription PCR, typical respiratory signs and symptoms, and chest radiographic findings compatible with pulmonary edema.

Heparinized whole blood collected by arm venipuncture was separated by centrifugation; plasma was stored at −20°C until analysis. In binding assays, antibody to all known hantaviruses indigenous to the Americas cross-react with the N protein of Sin Nombre virus (10). A strip immunoblot assay for IgG containing recombinant N protein of the 3H226 genotype of Sin Nombre virus was used as described (10); the criterion for positivity was a dark band for Sin Nombre N protein at a serum dilution of 1:200. An enzyme immunoassay used recombinant nucleocapsid protein from Sin Nombre virus (11); the cutoff value was established at 3 SD above reactivity to a panel of known positive serum. All samples were tested by both assays; the concordance of the enzyme immunoassay and strip immunoblot assay in this study was 97%, and the criterion for seropositivity was a positive reaction in both assays. Loss of antibody in persons with previously positive serum was determined by 2 independent tests with both assays. IgM against hantavirus was not tested. In Panama, all HPS patients tested have had positive reverse transcription PCR results for Choclo genomic RNA in acute-phase blood samples (9), and antibody has been detected by both assays.

Data were transferred from field collection forms to a database (Epi Info version 6.04d, Centers for Disease Control and Prevention, Atlanta, GA, USA) for statistical analyses using Epi Info software. Changes in seroprevalence within each community were tested by Fisher exact test for each interval and by longitudinal analysis for all intervals and communities by a generalized estimating equation (12). Increases in seroprevalence according to community and year of survey were tested by using analysis of covariance.

The 11 surveys repeatedly sampled 60%–85% of the total population of each community, for a total of 1,838 samples from 1,129 persons. Overall antibody prevalence was 32.9%, varying from 16.5% to 60.4% in individual surveys (Table 1). In each of the 3 Los Santos communities (Agua Buena, Isla Cañas, San Jose), seroprevalence increased annually by ≈5% (Table 1); the overall seroprevalence increases for the combined Los Santos communities were significant (Fisher exact test, p = 0.0014). The changes in seroprevalence were community specific (analysis of covariance F = 5.24, p = 0.0043), but increases in seroprevalence in the 4 communities combined was not (general estimating equation).

Among the study population, seroconversion was documented for 70 persons, and HPS was diagnosed for 5 other persons in the same communities during the intervals studied (Table 2). In the cohort of 476 persons in all 3 Los Santos communities sampled in 2 back-to-back surveys (Table 2), the 70 seroconversions occurred in persons in all age cohorts and equally among persons of both sexes (data not shown). No person who seroconverted gave a history of HPS-like illness or hospitalization for an acute respiratory illness. A separate study of outpatients with febrile illnesses was conducted in 4 clinics in the hantavirus-endemic area. This study found 48 adults and children with the typical febrile prodrome, unremarkable chest radiographs, and either serum IgM specific for hantavirus nucleoprotein or Choclo virus genomic RNA in the acute-phase blood sample (B. Armien and J.M. Pascale, unpub. data). These findings of symptomatic hantavirus infections confirm previous observations derived from neutralization-inhibition antibody assays (9).

The incidence of 70 seroconversions in 857 person-years of observation (Table 2) was equivalent to 8 infections per 100 person-years. The ratio of infection detected by seroconversion to infection resulting in HPS was 14:1. The mean ages of persons who seroconverted (43 years) and those with HPS (43 years) were the same. Undercounting of HPS cases was not likely because HPS is a highly publicized illness throughout Panama, and diagnostic serologic testing is readily available through the Ministry of Health.

A total of 16 seroreversions, compared with 70 seroconversions, occurred among persons in most age cohorts, mean age 48 years. For HPS caused by Sin Nombre and Andes viruses, serum antibody typically persists for years (13). Serum antibodies after mild or asymptomatic infections may not persist for many years.

Antibody prevalence surveys are useful for identifying populations and locations at risk, monitoring changes in incidence, and focusing limited public health resources. Determining whether the observed increases in seroprevalence will be sustained requires additional surveys, but this information will be useful as the new agroeconomy increasingly emphasizes the monoculture of products (rice and sugar cane) favorable to rodents (14). Nonetheless, the documentation of large numbers of mild or asymptomatic hantavirus infections not progressing to HPS has identified a larger effect of this zoonotic disease.