Institutional review board approval was obtained from Taiwan Disease Control, Ministry of Health and Welfare. Data of JE laboratory-confirmed cases between 1966 and 2012 were collected from the Taiwan Centers for Disease Control (CDC), and all the JE cases were anonymized and analyzed as secondary data. Serum samples were obtained from the Taiwanese Survey on the Prevalence of Hyperglycemia, Hyperlipidemia, and Hypertension (TwSHHH), funded by the Health Promotion Administration, Ministry of Health and Welfare, Taiwan in 2002. All the participants provided written informed consent before blood samples were collected, and all the serum samples were anonymized. In signing the written informed consent, all the TwSHHH participants also agreed that their remaining serum specimens could be used for other hyperglycemia, hyperlipidemia, and hypertension studies, as well as for public-health-related research.

JE has been a reportable communicable disease in Taiwan since 1955, and physicians are required to report incidents by using a hospital-based passive reporting system. Laboratory testing methods were established in 1965 to examine reported cases, followed by the implementation of an active JE surveillance system in 1967. Health care providers must notify the health authorities of any case that meets the case definition by law [13], and should collect and send specimens to the central CDC laboratory of Taiwan for examination. In addition, the local health authorities are responsible for epidemic investigation and related tasks, such as vector control. All suspected cases are archived and centrally managed by the Taiwan CDC.

According to a laboratory diagnosis, only two flaviviruses, JEV and the dengue virus, are known to be circulating in Taiwan. JEV is circulating nationwide, whereas the dengue virus is confined to Southern Taiwan. In 1998, the Taiwan CDC developed an E/M-specific capture IgM and IgG enzyme-linked immunosorbent assay (E/M-specific IgM/IgG ELISA) for JE and dengue fever (DF). However, antibodies against both JE and DF are screened using hemagglutination inhibition (HI) or an E/M-specific IgM/IgG ELISA, and the DF antibody test results are used as a negative control to reduce the probability of cross-reactivity. A long-term evaluation revealed that the E/M-specific IgM/IgG ELISA is highly sensitive and specific, and can effectively differentiate JEV from dengue virus infection [22]. Since 2001, acute phase serum obtained within 7 days and cerebrospinal fluid (CSF) of JE cases have been used for diagnosis through a real-time polymerase chain reaction (PCR) method; nevertheless, the E/M-specific IgM/IgG ELISA remains the primary method for diagnosis.

The JE laboratory confirmed that cases that met one of certain laboratory criteria were defined as being JE-confirmed cases. These criteria were: 1) an HI titer of the convalescent serum of ≥1∶160, and at least a four-fold rise between the HI titers of convalescent and acute sera; 2) an HI titer from a single serum of ≥1∶320; 3) an IgM positive serum was obtained using the ELISA test, or the IgG of the paired serum exhibited a four-fold increase; 4) a sample that exhibited a positive real-time PCR reaction; or 5) a sample that was positive to IFA staining after isolating the virus. In addition to meeting this case definition, JE cases are not confirmed until they have been positively diagnosed in a laboratory. This study comprised 2,948 cases out of the 2,954 cases confirmed between 1966 and 2012, excluding cases in six non-Taiwanese citizens.

The JE incidence rates of various age groups or birth cohorts were calculated using the 1966–2012 census population data obtained from the Department of Household Registration, Taiwan; these censuses are house-to-house and are conducted by the government every 10 years.

The Taiwanese government launched a comprehensive vaccination campaign against JE in the 1960s, during which all children younger than 3 years received two doses of JE vaccines. After 1974, the vaccination doses were increased to three with a booster dose administered one year after the two primary doses. After 1976, a fourth booster dose was administered to children during their first year at elementary school. After 1980, the target population for JE vaccination has mainly been children older than 15 months. The vaccination is administered in two doses separated by an interval of 2 weeks; it is administered before the epidemic season, which is from March to May, and this is followed by a booster dose one year later, and a final booster dose (the fourth dose) when the child enters the first year of elementary school.

The serum samples used were obtained from the TwSHHH (2002) [23], [24]. The TwSHHH subjects were enrolled from the National Health Interview Survey (NHIS) in 2001, which used a multistage, stratified, and clustered sampling scheme that included 26,685 noninstitutionalized residents from 6,592 households in 1,648 communities in Taiwan. As part of the TwSHHH, half of the NHIS-sampled Household Registration List (3,296 households) was randomly selected from each stratum. Out of 10,292 eligible subjects (73.6% attendance rate), 7,578 agreed to participate in the TwSHHH, and provided informed consent; 6,602 of the subjects who agreed to provide a blood sample were aged over 16 years, and 6,594 serum specimens were collected and included in the present study. Urbanization was divided into three levels, urban, suburban, and rural areas, according to previous studies [25], [26].

The plaque reduction neutralization test (PRNT) was used to detect the JE-neutralizing antibodies. The PRNT is the most frequently used and standardized method for detecting protective antibodies against the JE virus [27], [28]. In general, a PRNT50 titer of ≥1∶10 represents an effective protection [29], [30]. The JE vaccine strain that was used during the vaccination campaign was the Nakayama-NIH strain, and this strain was also used in the PRNT. The test sera were diluted (1∶10) in Dulbecco's Minimum Essential Medium (MEM) with 5% FCS after remaining inactive at 56°C for 30 minutes. An equal volume of JE virus solution, calculated to yield approximately 100 plaque-forming units (PFUs) per 0.1 mL, was added to each serum dilution, and incubated at 4°C overnight. The mixture was then added in duplicate to a monolayer of the cell line BHK-21 (Baby Hamster Kidney) in a 24-well plate. After adsorption for one hour at 37°C with 5% CO₂, the monolayer in each well was overlaid with 0.5 mL of 1% methyl cellulose in MEM containing 5% FCS. After 3–5 days of incubation at 37°C with 5% CO₂, the cells were stained with 0.1% Naphthol Blue Black, and washed with normal saline. The neutralizing antibody titer was pinpointed as the reciprocal of the serum dilution that inhibited 50% or more of the plaque formation compared with the virus control group.

The incidence rate of JE was calculated using the JE confirmed cases per 100,000 people at risk. A multiple logistic regression analysis was used to estimate the odds ratio and 95% confidence interval for each factor that was associated with the JE-neutralizing antibody. The dependent variable was the presence of a JE-neutralizing antibody in the individual: a value of 1 was assigned for the presence of the antibody, and 0 for the absence. The independent variables included sex, birth cohort (six cohorts represented by five dummy variables), and urbanity (three degrees represented by two dummy variables). The statistical significance level was set at 0.05. All computations were performed using the SAS 9.2 statistical software package (SAS Institute Inc., Cary, NC, USA).

The annual incidence rates and number of confirmed JE cases from 1966 to 2012 are presented in Figure 1. In this study, 2,948 JE-confirmed cases were analyzed. The incidence rates of confirmed cases dropped rapidly after 1971. The lowest incidence rate occurred between 1992 and 1997, with a slight increase from 1998, followed by a stable state.

The highest incidence rates, ranging from 1.65 to 2.04 cases per 100,000 people, occurred between 1966 and 1970. The incidence rates reached more than 2.0 cases per 100,000 people in 1967 and 1968. Approximately 200 to 300 confirmed cases were detected in these 2 years, and the number of cases substantially declined thereafter. The incidence rate decreased slowly and steadily from 1975 to 1991, ranging from 0.1 to 0.35 (0.233 on average) cases per 100,000 people, with approximately 15 to 67 confirmed cases every year. The lowest incidence rates corresponded to the period between 1992 and 1997, with 0.126 to 0.023 (0.066 on average) cases per 100,000 people, or between 5 and 27 confirmed cases every year. After 1998, the incidence rates were between 0.161 and 0.058 (0.128 on average) cases per 100,000 people, with an annual number of confirmed cases in the range of 13 to 37 cases, and 26 cases on average.

The average annual incidence rate of JE cases was approximately 0.118 cases per 100,000 people between 2002 and 2012 (Table 1). Only three of the 297 confirmed cases corresponded to vaccinated people. Two of them received only one dose, and the other received three doses; 60% of the confirmed cases were male, and 61.4% of them were older than 40 years.

The JE incidence rates of various age groups of the Taiwanese population are displayed in Figure 2. The epidemic status of JE over the years can be approximately divided into three stages. Before 1991, the confirmed JE cases primarily occurred in residents aged 0–29 years, with incidence rates that ranged between 0.124 and 3.03 cases per 100,000 people. The highest incidence rate in the age group of 0–29 years occurred between 1966 and 1970 (2.44–3.03 cases per 100,000 people), and declined rapidly thereafter. The average incidence rates and number of confirmed cases were similar for both age groups (<30 and ≧30 years old) from 1992 to 2000, and they were also the lowest compared to any other period. However, after 2001, the incidence rates among people aged 30 years or older were higher than those of younger people (0.167 vs. 0.052 cases per 100,000 people on average).

Figure 3 displays the incidence rates of JE across the three birth cohorts comprising people who were subjected to various vaccination schedules, namely those born before 1962, those born between 1963 and 1975, and those born after 1976. People born between 1963 and 1975 generally exhibit higher incidence rates than those of other groups, particularly during the period from 1967 to 1970 (from 5.4 to 6.82 cases per 100,000 people). The incidence rates were similar for the three birth cohorts from 1992 to 2002. Since 2004, the incidence rates of JE for those born between 1963 and 1975, or before 1962, have been slightly higher than those for people born after 1976.

The overall positive rate of JE-neutralizing antibodies in participants aged over 16 years was 71% in 2002. The age-specific seropositive rates of JE neutralization in people over 16 years in 2002 are displayed in Figure 4. The seropositive rate decreased gradually from 87% in people born in 1986 to 62% in people born in 1976 (who received four doses of the JE vaccine). The seropositive rates remained low (approximately 50%–60%) in people born between 1963 and 1975 (who received two or three doses of the JE vaccine). The seropositive rates increased gradually from 70% in people born in 1958 to approximately 90% in people born after 1943. People born before 1962 were never vaccinated because the immunization program had not yet been introduced.

To evaluate the influence of gender, birth cohort, and urbanization on the odds of presenting JE-neutralizing antibodies, a logistic regression model analysis was performed (Table 2). The results revealed that when the birth cohort and degree of urbanization were adjusted for, the odds of presenting JE-neutralizing antibodies were significantly higher in men than in women, with an odds ratio of 1.25 (95% CI = 1.12–1.40). Thus, the male JE seroprevalence was significantly higher than that of women.

The JE-neutralizing antibody positive rates were 74%, 63%, 55%, 54%, 68%, and 86% from the youngest to the oldest cohorts. When the gender and degree of urbanization were adjusted for, the birth cohorts of 1976–1980, 1970–75, 1963–1969, and 1953–1962 were less likely to test positive to JE-neutralizing antibodies than those born between 1981 and 1986 (aged 16–21 years); the odds ratios were 0.42 (95% CI: 0.33–0.53), 0.42 (95% CI: 0.34–0.52), 0.77 (95% CI: 0.63–0.94), and 0.56 (95% CI: 0.47–0.75), respectively. By contrast, people born before 1952 (aged over 50 years) were more likely to test positive than the youngest cohort; the odds ratio was 2.19 (95% CI: 1.78–2.69).

The JE-neutralizing antibody positive rates were 65%, 72%, and 76% for the urban, suburban, and rural groups, respectively. When gender and birth cohort were adjusted for, compared with people living in urban areas, people living in rural areas had an odds ratio of 1.61 (95% CI: 1.31–1.98) for testing positive to JE-neutralizing antibodies, whereas those living in suburban areas had an odds ratio of 1.52 (95% CI: 1.34–1.74) for testing positive to JE-neutralizing antibodies.