In 2000, 21 persons were reported with acute illness attributed to West Nile virus (WNV) infection; 19 were hospitalized with encephalitis or meningitis. Of the 21, 10 resided in the Staten Island borough (Richmond County) of New York City. Other ill persons resided in nine other counties—Kings (Brooklyn), New York (Manhattan), and Queens counties in New York; Hudson, Passaic, Monmouth, Morris, and Bergen counties in New Jersey; and Fairfield County in Connecticut. Because ill persons represent only a fraction of the persons who are infected, many more persons probably were infected in 2000. To determine the prevalence of recently acquired WNV infection and associated risk factors for infection, random household cluster serosurveys were conducted in Staten Island and portions of Fairfield County, Connecticut, and Suffolk County, New York, during October-November 2000. All three areas had intense WNV epizootics as determined by avian mortality and mosquito surveillance systems.1 This report summarizes the preliminary results of this survey and indicates that in areas with intense epizootic WNV activity, asymptomatic or mildly symptomatic human infections can occur.
After obtaining consent, persons aged ≥12 years were interviewed to identify risk factors for infection and submitted serum specimens for WNV IgM antibody testing. IgM-positive samples were tested for WNV neutralizing antibody.
A total of 2436 persons from 1989 (39%) of 5141 selected households participated in the serosurvey. Five persons aged 14-54 years had positive WNV IgM and neutralizing antibody tests indicating recent infection. Of 871 residents of Staten Island surveyed, four (unweighted seroprevalence estimate: 0.46%; 95% confidence interval [CI] = 0.18%-1.17%) had positive samples indicative of recent infection. Of 834 surveyed in Suffolk County, one (0.12%; 95% CI = 0.01%-0.67%) had a positive sample, and of 731 surveyed in Fairfield County, none (95% CI = 0.0-0.52%) had positive samples.
Of 176 persons reporting fever and headache during July-August 2000, two (1.1%) were infected recently, compared with three (0.1%) of 2222 persons who did not have these symptoms (relative risk = 8.6; 95% CI = 1.4-51.1; Fisher exact test, p = 0.05). However, persons recently infected with WNV did not differ significantly from other surveyed residents by age or sex.
In 2000, hospital-based surveillance identified 10 Staten Island residents with severe WNV neurologic illness (rate: 2.5 per 100,000 population). On the basis of Staten Island serosurvey data, an estimated 1574 (95% CI = 616-4003) residents aged ≥12 years were infected with WNV in 2000; an estimated one in 157 (95% CI = 1:62-1:400) WNV-infected Staten Island resident developed severe neurologic illness. In Suffolk County, although hospital-based surveillance did not identify any persons with severe WNV neurologic illness, an estimated 121 (95% CI = 10-673) infections occurred among the approximately 100,500 persons aged ≥12 years in the serosurvey area.
F Mostashari, MD, I Poshni, PhD, B Edwin, M Layton, MD, New York City Dept of Health; D Graham, MD, C Bradley, MD, Suffolk County Dept of Health Svcs, Hauppauge; M Kacica, MD, S Wong, PhD, C Franchell, MS, D Morse, MD, B Wallace, MD, P Smith, MD, State Epidemiologist, New York State Dept of Health. E Bresnitz, MD, State Epidemiologist, New Jersey Dept of Health and Senior Svcs. C Baisley, MPH, Greenwich Dept of Health, Greenwich; A Iton, MD, Stamford Dept of Health and Social Svcs, Stamford; G Archambault, MS, D Mayo, ScD, J Hadler, MD, State Epidemiologist, Connecticut Dept of Public Health. Arbovirus Diseases Br, Div of Vectorborne Infectious Diseases, National Center for Infectious Diseases; and EIS officers, CDC.
In 2000, the estimated incidence of recent WNV infection in three survey areas was less than the 2.6% estimated from a 1999 serosurvey in a north Queens neighborhood (CDC and New York City Department of Health, unpublished data, 1999). One possible reason for the lower incidence in Staten Island compared with Queens may have been that the 1999 WNV epizootic in Queens was more intense than that in Staten Island. Although few data exist to compare the epizootics in these boroughs, the seroprevalence of specific WNV neutralizing antibody among house sparrows was more than six times greater in north Queens in 1999 than in Staten Island in 2000. These differences may reflect the prevention measures implemented in 2000 that contributed to the decreased incidence in humans; these measures included mosquito larviciding before the transmission season, wide dissemination of public health messages promoting personal protection behaviors, reduction of peridomestic mosquito breeding sites, and intensive insecticide spraying to control adult mosquitoes. These differences also may reflect the sporadic nature of WNV outbreaks.2
Another important factor may have been the methods used to select the sites for the serosurveys. The 1999 Queens site was a 3 square mile area where the nine persons with severe WNV neurologic illness resided. Because the 10 case-patients in Staten Island were more evenly dispersed across the 56 square mile area, a sampling method that included the entire island was used. In 1999, the serosurvey results in Queens may have been lower if a wider area that included Queens neighborhoods with lower rates of severe neurologic illness had been used.
Fairfield and Suffolk counties were surveyed because of the many WNV-infected birds and mosquitoes reported. In Suffolk County, recent human infections were identified in the survey, although no cases of encephalitis were reported. In Fairfield County, although no recently infected persons were found, public health surveillance identified a mildly symptomatic resident with confirmed infection in 2000. The detection of WNV infection in these counties suggests that in areas with very intense epizootics human infections occurred but not at levels that resulted in recognized severe neurologic illness. Because older persons infected with WNV are more likely than younger persons to develop severe neurologic illness, in areas with equally intense epizootics and older residents, these lower infection incidences may still result in severe neurologic illness.
The comparable ratio of severe neurologic illness to infection observed in Queens in 1999 and Staten Island in 2000 suggests that, when adequate surveillance for severe WNV neurologic illness is in place, surveillance data may be used to estimate the WNV infection incidence from year to year. The identification of 62 persons in 1999 and 20 in 2000 with acute WNV illness suggests that approximately three times as many WNV infections occurred in 1999 as 2000 despite a widely expanding epizootic in 2000. Although some decrease in the rate of WNV infection in humans may be attributable to vector control, other prevention activities, or the variable and sporadic nature of WNV outbreaks, it is unknown why the estimated rates of infection varied widely among the three 2000 survey sites despite high levels of WNV epizootic activity in each. Additional analysis of the 2000 surveillance data will be necessary to identify surveillance indicators of increased risk for human infection to target prevention and control activities.
Serosurveys for West Nile Virus Infection—New York and Connecticut Counties, 2000. JAMA. 2001;285(6):727-728. doi:10.1001/jama.285.6.727-JWR0214-2-1