Lippincott Williams & Wilkins, Philadelphia, Pa. 5. obtained by a single-serum dilution (1:100), which correlated with ELISA titers obtained by an endpoint method. Under a tentative cutoff value (0.122) statistically calculated from NS1 antibody levels of horses in an area where JEV is not endemic, a high level of qualitative agreement (85.3%) was obtained between the ELISA and immunostaining methods. A significant correlation coefficient (0.799; < 0.001) was also obtained between the two methods. Three experimentally infected horses seroconverted no later than 13 to 23 days postinfection, whereas 4 field horses infected during an epizootic remained positive for NS1 antibodies for at least 40 weeks. Our results indicate that the ELISA used here was sufficiently sensitive to detect subclinical infections in vaccinated equine populations. Japanese encephalitis virus (JEV) is a member of the genus Flavivirus in the family (4). In a paradomestic environment, the virus is maintained in a transmission cycle between amplifier swine and vector mosquitoes (26). Humans are infected by bites from infectious mosquitoes and subsequently develop neurological diseases at subclinical:clinical ratios of 25:1 to 1 1,000:1 (31). This virus is distributed in many areas of Asia and parts of Oceania, with 50,000 human cases and 10,000 deaths per year reported (5). In Japan, several thousand human Japanese encephalitis (JE) cases were reported annually prior to the mid-1960s (8). However, a national JE vaccination program has successfully controlled the disease, keeping the annual number of confirmed human cases to fewer than 10 during the last decade (28). Horses are also susceptible to JEV. Although most infective mosquito bites result in subclinical infections, some infected horses develop fatal encephalitis, similar to humans. Large epizootics producing several thousands of equine JE cases in a single year were reported during the 1930s and 1940s (6). Following the introduction of inactivated JE vaccine in 1948, the number of reported cases decreased markedly (22). In recent years, racehorses, which have a substantial economic value, have been vaccinated every year prior to the epizootic season in Japan. No JE cases have been reported in racehorses since 1986 (19). The recent reduction in human and equine JE cases following vaccination indicates that vaccination has contributed 8-Hydroxyguanine to the control of JE (9). On the other hand, the reduction in JE cases might also have been the result, or at 8-Hydroxyguanine least in part, of less contact with infected mosquitoes (8). The use of insecticides and improved irrigation schemes for 8-Hydroxyguanine rice cultivation has reduced the number of vector mosquitoes. In addition, the structure and locations of pigpens have been changed so that pigs are now less frequently exposed to mosquito bites. As a result of these changes in Japan, Nfia concerns about side effects from the inactivated JE vaccine (1, 3, 23) have caused some to question the necessity of continuing vaccination. The most reliable strategy to address this question is to investigate how often humans and equines acquire natural JEV infections. Unfortunately, it has been difficult to obtain natural infection rates among vaccinated populations by conventional serologic methods, such as neutralization and hemagglutination-inhibiting (HAI) tests, since these methods are not capable of differentiating antibodies induced by natural infection from those induced by vaccination. A sensitive method based on immunochemical staining for the detection of antibodies to nonstructural 1 (NS1) protein of JEV has been established (14). Individuals vaccinated with inactivated JE vaccine consisting of viral structural proteins develop antibodies only to the structural proteins, whereas infection causes development of antibodies to nonstructural proteins in addition to antibodies to structural proteins. Therefore, detection of NS1 antibodies can identify naturally infected individuals among vaccinated populations. 8-Hydroxyguanine By this technique, a small-scale survey revealed relatively high annual infection rates in humans (5 to 10%) (14) and equines (15 to 67%) (13). Based on these results, it seemed desirable to carry out a larger-scale survey to show more clearly the situation in respect to exposure of humans and equines to natural infection with JEV. Enzyme-linked immunosorbent assay (ELISA) is simpler than immunostaining and more 8-Hydroxyguanine suitable for larger-scale surveys. However, ELISA has appeared less sensitive; while previous results showed that.