Background
Measles is a highly contagious disease caused by the rubeola virus, which has been in continuous global circulation and has significantly contributed to the worldwide disease burden. With the widespread use of measles vaccines (MVs), the incidence rate of measles has substantially decreased in most countries. Nevertheless, the disease still causes high morbidity and mortality among children worldwide, and 95% of measles-related deaths occur in low-income countries with weak healthcare infrastructures [
1].
China began deploying the MV in the 1960s and incorporated it into the routine immunization schedule under the national expanded immunization program in 1978, wherein a single dose of MV was administered free-of-charge to infants aged 8 months. Routine immunization with 2 MV doses commenced in 1985, with the first and second doses administered at the eighth month and the seventh year of age, respectively; the age of the second dose was lowered to 18–24 months in 2005 [
2]. In 2006, China launched the 2006–2012 National Action Plan for Measles Elimination, following which the incidence of measles was kept at a low rate mainly owing to enhanced routine immunization coupled with supplementary immunization activities (SIAs), which were mass immunization campaigns that targeted all individuals in a specific age range regardless of their immunization history [
3‐
5]. The purpose of the SIAs was to rapidly close the immunity gap within the targeted populations while providing herd immunity to newborns and older individuals. Unsynchronized province-wide SIAs involving MV administration were introduced in 27 provinces between 2003 and 2009 [
4], and a national synchronized SIA was implemented in 2010 [
3]. Although this program achieved great strides towards eradicating measles, a national resurgence of the disease occurred in 2013 along with changes in the epidemiological characteristics of measles infections [
3,
6‐
9].
Guangdong, a province in southern China, had a high and increasing measles incidence rate before 2009, with 21.2 per 100,000 population (the highest among all provinces) in 2007 and 19.1 per 100,000 population in 2008 [
10]. Despite the 2009 province-wide and 2010 nationwide SIAs, the number of measles cases began to surge in 2012 and was most notable in 2013 (accompanied by a national resurgence) [
11]. Guangdong had the highest measles incidence rate nationwide in both 2012 and 2013, and the second highest in 2014, and accounted for 30, 25 and 12.8% of the nationwide patients with measles in the three years respectively [
6,
7]. As the province with the largest population, Guangdong comprised 7.61% of the national population in 2011 and is critical to the country’s efforts towards eliminating measles [
12]. In this study, we aimed to investigate the changes in measles epidemiology during 2009–2016, particularly in terms of how the distribution of patients by age altered between the two periods of 2009–2011 (when the effects of the SIAs were in force) and 2012–2016 (when their effects diminished). Our findings can provide the epidemiological basis for improving immunization strategies for measles elimination in the province.
Discussion
The incidence of measles in Guangdong Province dropped to a very low level in 1–2 years after the SIAs were introduced but resurged afterwards; this was also the trend in other regions of China [
19‐
21]. Infants aged 0–8 months were the main group for the resurgence of measles, which was also the case nationwide [
6]. As indicated by some studies, this can be explained by the 0–8-month-old patients being the most susceptible to measles outbreaks among all age groups. Although infants had immunity acquired from their mothers, most had this immunity waned in 6–8 month.
A sero-survey performed in Guangdong in 2013 revealed that only 23.58% of infants who were 4–5 months old were seropositive, as were only 10.53% of those 6–7 months old and 54.17% of those 8–9 months old [
22]. Another serological study conducted in Guangdong found that infants as young as three months of age were sometimes seronegative if their mothers had low levels of measles antibodies [
23]. Even women who have acquired natural immunity or been vaccinated may not transmit sufficient amounts of maternal antibodies to protect their infants during their first eight months of life. Based on these data, the majority of 6–8-month-old infants appear to be particularly susceptible to measles (i.e., have a “zero defense” status) [
22,
24]. Moreover, the proportion of unprotected infants is high because mothers nowadays mostly acquire their immunity from vaccinations (especially those born after 1985 when 2-dose MVs were introduced) rather than from natural infections; hence, infants have less transmitted immunity with shorter protective durations [
25,
26]. Nevertheless, a Belgian study in a nearly eliminated setting showed that maternal immunity actually waned at an early age in the children of both vaccinated and unvaccinated mothers [
27,
28]; this warrants further investigation.
We conducted a further analysis of children who were too young for routine immunization (See Additional file
1). Those who were six to eight months of age comprised nearly 80% of all infants ≤8 months infected annually; a large percentage had therefore lost maternally transmitted immunity. Every group had almost the same proportion of cases during 2012–2016 as they did in 2009–2010; during the 2013–2014 resurgence, there was no increase in the proportion of infants < 6 month of age who were infected. Our evidence cannot draw conclusion on the relationship between resurgence and the trend of waning immunity, due to the lack of serological data, but the large proportion of 6–8 months old cases had indicated the need to pay attention to the immunity gap between when infants’ maternal immunity waned and vaccination administration at the eighth month. China’s immunization schedule administers MV1 at eight months of age, which is earlier than the World Health Organization’s recommendation (nine months) [
29]. In general, maternal immunity that has not yet waned may interfere with the vaccine-induced immune response and result in vaccination failure [
30]. Since most infants in Guangdong who are eight months have already experienced a waning in their immunity [
22], 94–96% of them can successfully be seroconverted after vaccination [
31]. However, advancing the MV1 delivery time to 6 months of age remains controversial [
7], even though the World Health Organization suggests that infants aged ≥6 months should have a supplementary MV before routine immunization if they are in high-risk settings like daycare facilities [
29]. Aside from closing the immunity gap among 6–8-month-olds, protecting unvaccinated children via herd immunity and preventing transmission from older populations should be a primary focus [
7].
The second reason for the surge of measles in 0–8-month-olds was the increased number of infective individuals transmitting the virus to susceptible, unvaccinated infants. The increased infection rate among 9–23-month-olds and 2–6-year-olds were due to the lower coverage in birth cohorts after the 2010 SIAs as well as the absence of extra doses offered by the SIAs [
32]. At the same time, a multi-provincial study revealed that one of the most prevalent routes of exposure was hospital visits [
33]. A case-control study in Guangdong province also showed that children who visited hospitals had an odds ratio of 5.50 of developing measles within one to three weeks compared to those who did not; this was the second most significant factor after non-vaccination [
34]. Moreover, infants are prone to other illness and might visit healthcare facilities more frequently for medications and regular check-ups; thus, infection control in healthcare settings is crucial for reducing the incidence of measles. In our dataset, we found that 1136 infants diagnosed with measles during 2012–2013 who were 0–8 months old (40.6%) had visited hospitals 7–21 days before disease onset, while 1478 (52.8%) had not and 183 (6.5%) had an unknown hospital visit history. The proportion of patients having been exposed to hospital settings was slightly lower than that shown in the aforementioned multi-provincial study (45%), with outpatient hospital visits being a significant predictor of infection (adjusted matched odds ratio 9.4) [
33].
The rise in measles incidence rates among older age groups, especially the 9–23-month-old and 2–6-year-old groups, resulted from the aggregation of susceptible individuals after the province-wide SIAs were introduced, owing to lower vaccination coverage or primary vaccine failures. Low vaccination coverage among newly born cohorts after the 2010 SIAs was a key contributor to the national resurgence [
32,
35]. A recent survey conducted in Guangdong revealed that MV1 coverage was 71.8% among migrant children aged 12–59 months, and that only 37.2% had MV1 administered in the appropriate time [
36]. On the other hand, increased vaccination coverage among children ≥8 months had very positive effects on providing herd immunity to those ≤8 months. Owing to the SIAs in 2009 and 2010, incidence in every age group dropped by 60–85% in 2010; interestingly, only the 0–8- and 9–23-month-old groups had similar and substantial drops in the number of cases in 2011 as they did in 2010 (i.e., decreases of 70 and 60% respectively).
Comparing across age groups, we showed that the proportions of patients with measles aged 7–15 years and 16–24 years had dropped significantly between 2009 and 2011 and 2012–2016, as they were covered by the SIAs. Proportion of 7–15 and 16–24 years old cases were higher during 2009–2011 (9.9 and 17.5% respectively) but lower during 2012–2014 (4.0 and 7.5% respectively), indicating the effectiveness of SIAs in protecting the cohorts against outbreaks. However, similar to other Chinese provinces at the time, Guangdong experienced increasing proportions of adults with the disease from 2014 to 2016; adults accounted for the highest number of diagnoses in 2016. Relatively high proportions of adult patients were also observed after the SIAs were introduced in other studies [
20,
21]. The decreasing incidence in the overall population, declining birth rate, and increasing vaccination coverage among infants and children compared to the low coverage rate in the 1980s and earlier [
37,
38] all contributed to the increasing mean age of infection and build-up of susceptible individuals among adults [
39]. In China, infections among adults were mainly due to susceptibility caused by missed vaccinations and decreased opportunities for acquiring immunity through natural infection [
38,
40]. As revealed in our study and others, the majority of infected adults were either unvaccinated or had an unknown vaccination history [
41,
42].
Various vaccination strategies have been suggested for reducing the number of susceptible individuals among infants, eligible children, and adults. First, for infants with waned maternal immunity, MV1 was not recommended at any time earlier than eight months of age on a routine immunization basis, as infants with maternal immunity that did not wane below a certain threshold would experience vaccine failure [
43,
44]. An additional MV is suggested for infants six months of age in cases of outbreaks; these should be followed by MV1 and MV2 per the regular schedule [
29,
45]. Second, for 8-month- and 18–24-month-old children eligible for MV1 and MV2, respectively, responsible authorities should maintain a high coverage and timely delivery of routine immunization through frequent monitoring and coverage assessment [
32]. Lastly, investigations on adult susceptibility, especially in terms of secondary vaccine failure, were needed, and procedures to reduce susceptibility among adults might be warranted [
42]. Although the World Health Organization and some studies have highlighted the successes of adult-targeted SIAs [
46,
47], mass immunization campaigns targeting adult susceptibility were not recommended because selective and targeted SIAs; i.e., revaccination of school students, women of child-bearing ages, and populations susceptible to outbreaks, were rightfully deemed more appropriate and useful than non-selective SIAs [
40,
42].
On case distribution by cities during 2013 outbreak, Guangzhou and Shenzhen had high number of cases because the two cities hosted 21.6% of total population in Guangdong and had very high proportions of migrants population [
48], among whom incidence was higher and age-appropriate MCV uptake rate was low [
36,
49]. As for Huizhou, a city with lower population density and fewer migrants [
48], incidence among MCV-eligible children was very high; 9–11 months old cases alone accounted for 24.8% of total cases in 2013 [
50]. In Zhanjiang city with low population density and a net outflow of population, incidence among the 12–23 months old was as high as 122.9 per 100,000 in 2013 [
51]. The two cities also had higher percentages of cases with no or unknown immunization history (89.8% in Zhanjiang overall and 88.2% in Huizhou among 8 months – 14 years old cases) compared to Guangdong. Missing or delayed routine immunization is yet to be addressed so as to control measles in the two cities.
There were several limitations in this study. First is the possible underestimation of measles cases because some patients might not present to local doctors. Second is the limited analysis of the effectiveness of the two SIAs owing to the lack of measles data for several years before the SIAs were introduced in 2009. In our study, we focused on newly born cohorts after the SIAs (0–23-month-olds in and after 2012). Unlike those who were born in 2009–2010 and were protected by the herd immunity provided by older age groups (the 8 month- to 14-year-olds in 2009 and the 8 month- to 4 year-olds in 2010, respectively), newborns in 2012 or later had older (≥8-month-old) counterparts who did not receive additional vaccine doses administered through the SIAs. Therefore, the effectiveness of SIAs in terms of influencing the incidence of measles among newborns born within or outside their implementation periods could be assessed by examining the changes in epidemiological characteristics (e.g. age distribution) during the post-SIA resurgence in 2013. Third, the reported vaccination coverage periods for the two SIAs were only administrative estimates, and were likely slightly overestimated. Our main focus was on how newborns were protected by the herd immunity of different cohorts whether subject to the SIAs or not, and we posit that the overestimation of the administrative coverages would not appreciably affect the conclusions of our analysis since the total vaccinated population still accounted for an overwhelming majority of the population (e.g., ≥19.9 million vaccinated individuals in 2009 accounted for 93.9% of population in the eligible age range). Fourth, the routine immunization coverages during the study period were not provided to correlate with the age-specific incidence, due to possible over-estimation. Susceptibility among different age groups were discussed in correspondence with the different vaccination strategies targeting every individual in specific age groups (e.g. routine immunization for the 8 and the 18–24 months old, SIAs for the 8 months old to 14 years old in 2009), and therefore the trend of incidence in different age groups reflected effectiveness of those interventions in different age groups. Fifth, the vaccination history of the majority of individuals ≥16 years of age were unknown, although the percentage of unknown vaccination history was small among young children between 2012 and 2016; this limited our analysis on vaccination history. Lastly, the study was of limited statistical power owing to its secondary data analysis design that lacked a hypothesis-driven sample size planning.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.