According to the Western Pacific Regional Plan of Action for Measles Elimination (WHO, January 2003), the elimination of measles can only be achieved if population susceptibility is below the level that can sustain transmission of measles. Thus, 95% population immunity is necessary to interrupt transmission and eliminate measles 
. Our 3 years of successive cross-sectional surveys revealed that the overall measles seroprevalence was 91.7% in 2010 in the Jiangsu province; this needs to be further increased to achieve the elimination goal. When considering the immunity of the population older than 12 months in age, the seropositive rates for the 3 surveyed years were all over 95%, but there were still wild measles viruses circulating among individuals of these ages. These data suggest that an overall seropositive rate in the province higher than 95% may be insufficient for elimination of measles. For example, the seropositive rate of those older than 12 months was 95.3% in 2010; however, the rate was 88.4% in adults aged 20–29 years old, which may be representative of an epidemic and outbreak of measles due to the frequent migration of adults. Therefore, one of the operational indicators for measles elimination is to maintain 95% immunity to measles in each cohort in every district 
. Strategies may vary according to population immunity in each age group in each city or county.
The Pan American Health Organization has interrupted indigenous measles transmission since November 2002. One of the primary strategies used to eliminate measles was the strengthening of routine vaccination 
. In contrast, the experiences of some European countries showed that SIAs were essential to measles elimination 
. Our survey supports SIA-based strategy. For instance, although the reported coverage with routine two-dose measles vaccine was over 95% in most regions of the province in 2008, the seroprevalence in several age groups from our survey was lower than 95%, indicating a high risk of measles epidemics to the general population in the province 
. Further evidence supporting this hypothesis was the significantly higher incidence of measles in the last 2 months of 2008 when compared to 2007.
In February 2009, catch-up SIAs were performed across the whole province among children between 8 months to 15 years old. A total of 9,679,489 children were vaccinated within 2 months, and vaccination coverage of 97.2% was achieved. To accelerate measles elimination and based on the risk assessment of measles epidemics and immunity in the general population, follow-up SIAs in children aged 8 months - 4 years were conducted in October 2010. In that SIA, 4,254,194 children were vaccinated with 95.3% coverage. The most important impacts of SIAs were the upward trend in the immunity of the general population shown in this study, where the measles incidence was observed to dramatically decline from 73.7 per million in 2008 to 9.5 per million in 2010 following the SIA. Several studies also indicated that the estimated reduction in the population susceptible to measles was more than 75% after SIAs 
The purpose of a follow-up campaign is to reduce any accumulation of susceptible individuals born since the previous SIA. These are conducted every 2 - 4 years following catch-up SIAs 
. The interval between campaigns was only 20 months in Jiangsu province because (1) both rapid, convenient monitoring and coverage surveys after the campaign indicated that the coverage rates in 3 of 13 cities were still lower than 95% in the catch-up SIAs, and (2) the seroprevalence of children from 8 months -15 years old in the central region was lower in 2009 when compared to 2008. The results of the serological survey after the 2010 follow-up campaign revealed that the seroprevalence of 8 month - 4 year olds was significantly higher when compared to 2009 in the central region. In contrast, a small rise in GMT and relatively stable seroprevalence were seen after a follow-up campaign in the other regions.
When the results of these surveys were interpreted longitudinally, seroprevalence and GMTs in children aged 8–12 months (when the first vaccination was offered) was lower than 83% and 1.00 IU/ml, respectively. In children 25 months - 4 years old (when the second vaccination was offered), a significant rise in the prevalence of protective levels of measles antibodies and GMT was observed. These data demonstrated that a two-dose schedule can provide a second chance at immunization for primary vaccine failures and for persons who missed the first dose. The results from other studies showed that short-lived antibodies titers increased after booster immunization with live attenuated vaccine 
, which may represent the immune response when individuals with pre-existing immunity are vaccinated.
Another challenge to the elimination of measles is that, despite the very high seroprevalence among children aged 24 months to 14 years old, the GMTs showed a downward trend with age (). This could imply that vaccine-induced measles antibodies wane in time, as indicated by previous studies 
. Both GMTs and seroprevalence declined in adults from 15–29-year-olds who were born in the post-vaccine era. Gaston De Serres and others recently found indications that there may be a higher risk of measles among high school students when the first dose of measles vaccine is given before 15 months of age 
Our study suggests that administering the first dose of measles vaccine at 8 months of age may not be the optimal age at this stage in the elimination of measles. China provides the first dose of measles vaccine at the young age of 8 months. In 2009, the WHO recommended that countries nearing the elimination of measles consider administering the first dose of measles vaccine at 12 months of age. However, prior to changing the age of the first dose of measles vaccine, the WHO recommends that 4 pieces of evidence be considered: (1) the seroconversion rate by age of vaccination, (2) the actual ages of administration for the first dose, (3) projected differences between vaccination rates at 8 and 12 months of age, and (4) the age-specific incidence of measles 
Considering all 4 lines of evidence is essential to making an informed policy decision. For example, with more mothers having vaccine-induced antibody levels, the proportions of their children who derive maternally protective levels of antibodies may be lower in the future 
. Infants will be at a higher risk of being infected before vaccination if measles circulates. In line with our studies, surveillance over the whole province found increasing proportions of infants younger than 8 months old who were infected with measles in recent years.
The incidence of measles in adults aged 20–29 years was also high, at approximately 70–100 per million from 2005–2010 
. Some studies have revealed that conducting a revaccination program in secondary school may be a way to improve immunity for adolescents, which may improve the interruption of measles virus circulation in young adults 
. Most adults aged 30 years and older were born in the pre-vaccine era and thus infected measles. Persistence of protective measles antibodies after natural infection seems to be for life