Dynamic epidemiological changes of hand, foot, and mouth disease and real-world effectiveness of EV-A71 vaccination: A case study in Shanghai (2009–2023)

ABSTRACT

A significant reduction in hand-foot-mouth disease (HFMD) cases has occurred across the country, since EV-A71 vaccine was licensed for use in China in 2016. We compared the epidemiology and virological characteristics before and after the implementation of EV-A71 vaccination over a 15-y period in Minhang district, Shanghai, and conducted a test-negative design to estimate the vaccine effectiveness (VE) of EV-A71 vaccine against HFMD. A total of 73,160 HFMD cases, 361 severe cases, and 3 fatalities were reported in the Minhang district, Shanghai during 2009–2023. After the implementation of EV-A71 vaccine program, the incidence rate, case-severity rate, and fatality rate of HFMD showed a significant decline by 56.7%, 95.2%, and 100.0%, respectively. The proportion of age group 6–10 y increased by 105.7%. The predominance of EV-A71 and CV-A16 was replaced by CV-A6 and CV-A10 in the post-2017 period. Full-dose immunization coverage rate maintained 50% during 2019–2023. The overall VE against EV-A71-associated HFMD and herpangina was 89.9% (95% CI: 74.5–96.0) for two-dose series and 67.9% (95% CI: 2.6–89.6) for one-dose vaccination. The VE for two-dose vaccination was 92.2%, 88.1%, and 100% against EV-A71-associated outpatient visit, non-severe hospitalization, and severe complications, respectively. There was no cross-protective effect of EV-A71 vaccination against non-EV-A71 serotype infection. EV-A71 vaccination program substantially reduced HFMD burden in Shanghai, with two-dose series providing robust protection against EV-A71-associated HFMD and herpangina. The shift of non-EV-A71 serotypes warrants an effort to develop multivalent vaccines to control HFMD epidemics.

Introduction

Hand-foot-mouth disease (HFMD) is a common childhood infectious disease caused by Enterovirus alphacoxsakie.¹ In the past two decades, the large outbreaks of HFMD constantly occurred in the Asia–Pacific region,^2–6 resulting in life-threatening neurological complications and fatality predominantly in children under 3 y of age. Enterovirus 71 (EV-A71) was the primary causative agent responsible for the outbreak and severe outcomes.⁷ HFMD was included to the legally notifiable infectious diseases, soon after the nationwide large outbreak of EV-A71-associated HFMD triggered a pressing public health concern in 2008 in China. The annual large epidemic of HFMD was curbed until the first inactivated EV-A71 vaccines were available in China since 2016. The EV-A71 vaccines were first licensed in December 2015 in China and approved for use in children aged 6–59 months, with two-dose series administered 28 d apart. The Phase III clinical trials showed that vaccine efficacy was above 90.0% against EV-A71-associated HFMD and 80.4% against EV-A71-associated disease.^8–10 Although EV-A71 vaccination is self-paid by family, the parents’ willingness to vaccinate their child was relatively high (~50%) in China because most of parents concerned about the harm of HFMD and trusted the benefit of EV-A71 vaccines.^11,12

Shanghai is one of the largest megacities on the eastern coastline of China with nearly one half of migrant population. Prior to the implementation of Shanghai EV-A71 vaccination, the average annual incidence of HFMD in Shanghai was much higher than the national level (179/100,000 vs. 135/100,0000) and ranked second in Eastern China during 2008–2017.¹³ The economic burden of HFMD in Shanghai reached over US$7.66 million in 2011, of which 46.1% was attributed to EV-A71 infection.¹⁴ Despite annual outbreaks of HFMD since 2008, the antibody level of EV-A71 was 35–40% in Shanghainese children aged 1–5 y during the period of 2012–2016.¹⁵ The successful introduction of EV-A71 vaccination in China plays an important role in preventing severe HFMD and controlling the annual outbreak of HFMD.^8,9,16 The real-world studies conducted in other provinces, such as Henan,¹⁷ Beijing,¹⁸ and Guangxi¹⁹ showed the estimated effectiveness of EV-A71 vaccine against EV-A71-associated HFMD ranged from 83.7% to 88.3%, which was derived from single outpatient visits and hospitalizations during 1-y observation period. To date, the long-term effectiveness of the EV-A71 vaccines has not been reported. Therefore, we carried out a real-world study to assess the effectiveness of EV-A71 vaccine among children and the effect of EV-A71 vaccination program on the epidemiology of HFMD in Shanghai based on the surveillance data of HFMD from 2009 to 2023.

Methods

Study site and data collection of HFMD

This study was carried out in Minhang district in which there is a tertiary children’ hospital (Children’s Hospital of Fudan University, CHFU) that is a sentinel surveillance hospital of notifiable infectious disease and is designated specifically to treat notifiable infectious diseases. Most inpatients with HFMD and herpangina and all severe cases are referred to this hospital for management in Shanghai. Minhang district is the second largest district of Shanghai with an area of 370.75 km² and 2,650,000 inhabitants in 2020. It is a typical area for the inflow of migrant population as well, with 1,460,000 migrant inhabitants in 2020.

Data on HFMD cases between January 2009 and December 2023 were obtained from the National Notifiable Diseases Reporting System (NNDRS). Information on the notified cases included case number, date of symptom onset, clinical severity, and demographic characteristics, such as gender, birth date, occupation, household registration, and residential address. All legally notifiable HFMD cases and herpangina (nonstatutory notifiable infectious disease in China) cases were clinically diagnosed based on the criteria of national guideline.²⁰ The severe case is defined as the clinically diagnosed patient accompanied by central nervous system involvement, such as meningitis, encephalitis, and acute flaccid paralysis (AFP), as well as neurogenic shock, and cardiorespiratory failure or pulmonary edema secondary to brainstem encephalitis, and life-threatening myocarditis.^20,21

Pathogen surveillance

The CHFU is responsible for routine pathogen monitoring of HFMD and herpangina in Shanghai since 2010. Over 90% of outpatients and 100% of inpatients and severe cases of HFMD reported in Minhang district are treated at the CHFU. As per the surveillance protocol, the 3–8 pharyngeal swabs were collected each week from the first 3–8 HFMD/herpangina outpatients upon visiting the hospital, the stool samples were routinely collected from the HFMD/herpangina inpatients if stool available during hospitalization, cerebral spinal fluid (CSF) was collected from severe cases complicated with meningitis, encephalitis, and AFP. The corresponding demographic and clinical information of every sampled patient was obtained from the hospital electronical information system and EV-A71 vaccination history was traced through Electronic Immunization Information System in Shanghai. All samples were tested for pan-enterovirus, EV-A71, CV-A16 using real-time quantitative PCR (qPCR) with detection kits (Shanghai BoJie Technologies Co., Ltd., CHN), according to the manufacturer’s instructions. CV-A6 and CV-A10 were added to the enterovirus panel since 2013.

Study population

The total population of Minhang district, encompassing both local registered residents and migrant residents, was extracted from the Shanghai Population Statistical Yearbook. In this study, the migrant residents refer to individuals who are away from their usual place of residence and reside in Shanghai for a minimum of six consecutive months. The number of children aged 0–5 y, including both local registered and migrant populations, was retrived from the Shanghai Management System of Information for the Immunization Platform (SMSIIP) which is managed by 14 uniformly trained medical staff at Community Health Service Centers across Minhang district. For migrant children, they are registered to SMSIIP if they get vaccination in Minghang district for at least 3 months after vaccination. Notably, vaccination services in Shanghai are equally accessible to both local registered and migrant families, ensuring that every child has the opportunity of getting vaccination either in their place of birth or their current city of residence. This equitable service coverage ensures the reliability of the data on the number of migrant children aged 0–5 y included in this study.

Evaluation of the EV-71 vaccination impact

To evaluate the impact of vaccine introduction on the epidemiology of hand, foot, and mouth disease (HFMD), the study period was stratified into two distinct stages: the pre-vaccination era (2009–2016) and the post-vaccination era (2017–2023). We calculated the epidemiological indicators (incidence rates, severity rates, and mortality rates), analyzed the proportions of cases across different demographic characteristics (gender, age, and occupation), described the seasonal patterns (the epidemic curves over 15 y and by 12 months), and compared the ratios of predominant pathogen trends for HFMD of different severities between the two stages.

Real-world evaluation of EV-A71 vaccine effectiveness

The EV-A71 vaccination was introduced in Shanghai on October 31, 2016. To evaluate the vaccine effectiveness (VE) of EV-A71 vaccination against EV-A71-associated HFMD and herpangina, we conducted an unmatched case-control study. Data were from January 1, 2017, through December 31, 2023, based on the HFMD/herpangina virological surveillance system in Minhang district, Shanghai. Given the limited sample size, an unmatched case-control design was employed, with multivariable regression analysis subsequently performed to adjust for potential confounding factors. Our study included mild outpatients, nonsevere inpatients, and severe inpatients diagnosed with HFMD/herpangina from the surveillance hospital. Asymptomatic or subclinical EV-A71 infections were excluded, as such cases are unlikely to seek medical attention, undergo laboratory testing, or be captured by mandatorily surveillance systems. The inclusion criteria for case and control were as followings: (1) children aged 6–59 months (age-eligible for EV-A71 vaccination), (2) children resided in Shanghai, (3) children with a definite EV-A71 vaccination history, (4) children with PCR-confirmed enterovirus infection in the clinical samples. Case was defined as a virological surveillance case who was tested positive for EV-A71 and pan-enterovirus, while control was defined as a virological surveillance case who was tested positive for pan-enterovirus but negative for EV-A71. For enrolled inpatients, a diagnosis of enterovirus positivity was made if either pharyngeal swabs or fecal specimens tested positive, which enhanced the sensitivity of pathogen detection and minimized misclassification bias.

Shanghai Management System of Information for the Immunization Platform (SMSIIP) was established since 2011. For the local Shanghai household children, the EV-A71 vaccination record can be reviewed through the platform. For the migrant children whose households were from the outside of Shanghai, the evidence of EV-A71 vaccination was collected from written vaccination record or parents or guardians authentic recall. Unknown or uncertain immunization status was considered missing vaccination. EV-A71 vaccination was defined as a child receiving one dose or two doses of EV-A71 vaccine at least 28 d apart.

In the test-negative study, VE was estimated as 100% × (1 − OR_adj), where OR_adj was adjusted odds ratios of EV-A71 vaccination history and EV-A71 detection and calculated in the conditional logistic regression model adjusted for gender, age group (6–35 months, 36–59 months), clinical severity (mild outpatient, nonsevere inpatient, and severe inpatient), clinical diagnosis (HFMD, herpangina), and matched on year of onset (2017–2019, 2020–2023) to account for changers in vaccination coverage and risk of infection over time. The 95% confidence interval (CI) for vaccine effectiveness (VE) was calculated using the following formula: 95%CI = VE ± 1.96 × SE, where SE was the standard error of OR_adj from the logistic regression. We used the Wald method to calculate the confidence interval. The 95% confidence interval means that 95% of the calculated intervals would contain the true population parameter value of VE underrepeated sampling.

We also estimated the potential nonspecific VE for EV-A71 vaccination against CV-A6, CV-A16, and CV-A10 associated HFMD/herpangina, respectively, using the same test-negative design and logistic model to determine whether there might be nonspecific effectiveness of vaccination. The VE analysis against CV-A16, CV-A6, and CV-A10 were based on RT-PCR testing results of throat swabs and stool combined, with those testing positive for CV-A16, CV-A6, and CV-A10 as test-positive group and those testing negative for CV-A16, CV-A6, and CV-A10 as test-negative group, but excluding those testing positive for EV-A71. Data were analyzed with SPSS 26.0 software (IBM Corp, USA). The characteristics between different groups were compared using χ² tests or Fisher’s exact test as appropriate. All statistical tests were two-sided, and statistical significance was defined as p < .05.

Results

Comparison of incidence and demographic changes before and after vaccination

A total of 73,160 HFMD cases, 361 severe illness with complications, and 3 deaths were reported to China CDC during 15-y monitoring, with an average incidence rate of 193.8/100,000, case-severity rate of 0.49%, and fatality rate of 4.1/100,000 in Minhang district. After the implementation of the EV-A71 vaccination, the average incidence rate of HFMD decreased significantly from 267.6 per 100,000 population in the prevaccination period to 115.9 per 100,000 population in the postvaccination period, with a reduction of 56.7%. The case-severity rate and the fatality rate from 2017 to 2023 have also declined sharply, 95.2% and 100.0%, respectively.

From 2009 to 2023, the male-to-female ratio of HFMD patients was 1.49:1 (ranging from 1.45:1 to 1.51:1). Of all the reported cases of HFMD, 0–5 y old children accounted for 87.70%, and 55.24% of them were home-care toddlers (0–3 y old). Since the period of post-EV-A71 vaccination, the proportion of 0–5-y-old children and home-care toddlers declined by 13.23% and 17.23%, respectively, while the proportion of 6–10-y-old children and school-age children (6–18 y old) increased by 105.68% and 202.69%, respectively (Table 1).

Table 1.

Comparison of HFMD demographic characteristics before and after the EV71 vaccination in Shanghai Minhang district, 2009–2023.

	Total (N = 73,160)		2009–2016 (prevaccination) (N = 51,885)		2017–2023 (postvaccination) (N = 21,275)		X2	P-value
	Number of cases	Ratio (%)	Number of cases	Ratio (%)	Number of cases	Ratio (%)
Gender
Male	43,812	59.89	31,237	60.2	12,575	59.11	0.21	.885
Female	29,348	40.11	20,648	39.8	8700	40.89
Age
0 –5	64,159	87.70	47,322	91.21	16,837	79.14	2218.46	<.001
6–10	6906	9.44	3747	7.22	3159	14.85
>10	2095	2.86	816	1.57	1279	6.01
Occupation
Home-care children	40,413	55.24	30,173	58.15	10,240	48.13	2490.17	<.001
kindergarten children	25,902	35.41	18,618	35.88	7284	34.24
School children	6044	8.26	2696	5.2	3348	15.74
Adults	801	1.09	398	0.77	403	1.89
Total	73,160	100	51,885	100	21,275	100

Temporal and seasonal trends

The epidemic wave of HFMD fluctuated every 2 y. The annual incidence rate of HFMD showed a rising trend year by year from 195.9/100,000 in 2009 to an incidence peak of 373.8/100,000 in 2014, then decreased to the nadir (10.60/100,000) in 2022 but sharply rebounded to 165.3/100,000 in 2023 (Figure 1).

Figure 1.

Aunnal number of HFMD cases and incidence rates of HFMD in Shanghai Minhang district, 2009–2023.

The biannual peak seasons of HFMD in Shanghai were seen in the summer (May to July) and autumn–early winter (September to December). Each monthly mean incidence rate in the postvaccination period is lower than that in the prevaccination period, except for September. After the introduction of the EV-A71 vaccine, the summer peak delayed from June to July, and the winter peak didn’t appear (Supplementary Figure S1).

Distribution of enterovirus serotypes

A total of 2653 mild outpatients and 331 severe inpatients of HFMD were included to virological monitoring from 2009 to 2023 in Minhang district, Shanghai.

Among the mild outpatients, the percentages of EV-A71 serotype decreased dramatically from 32.0% (2009–2015) to 3.6% (2016–2023) (x²  = 422.5, p < .001) after the use of EV-A71 vaccine. EV-A71 accounted for the highest percentages of 74.8% in 2011, then sustained a stable plateau level with annual percentage ranging from 19.3% in 2012 to 29.4% in 2017. The percentage of EV-A71 declined to the nadir level (1.8%) in 2018 and was no longer detected in mild cases since 2019. CV-A16 was most predominantly prevalent in 2010 (39.4%), 2012 (57.4%), 2014 (38.2%), and 2019 (39.3%), with predominance every 2–5 y and was not commonly detected since 2020. CV-A6 replaced the predominance of EV-A71 and CV-A16 since 2015, ever accounted for 50.0–95.9%. The prevalence of CV-A10 was always at a low level of less than 8.0%. Other enteroviruses than EV-A71, CV-A16, CV-A6, and CV-A10 ever accounted for 41.5% in 2013 but maintained a prevalence rate of 0.9–13.5% since 2014. There is a linear correlation between the prevalence intensity of EV-A71 and severity incidence rate of HFMD (Figure 2(A)).

Figure 2.

The distribution of enterovirus serotypes among HFMD cases by different severity. (A) The annual distribution of enterovirus serotypes among mild HFMD cases and the severity incidence rate of HFMD in Shanghai Minhang, 2010–2023. (B) The annual distribution of enterovirus serotypes and the total number of cases among severe HFMD cases in Shanghai Minhang district, 2009–2023.

EV-A71 accounted for 81.3% to 100.0% in severe cases from 2009 to 2017. With the increased uptake (39%) of EV-A71 vaccine since 2018, severe HFMD cases reduced substantially by 97.9% and no severe case of EV-A71 infection was identified since 2018. A few of sporadic severe cases were associated with CV-A6, CV-A16, and nonserotyped EV, which accounted for 0.3%, 1.4%, and 4.8%, respectively, from 2009 to 2019. No severe case was reported after 2020 (Figure 2(B)).

Evaluation of EV-A71 vaccine effectiveness

The number of EV-A71 vaccine doses administered to children <6 y of age increased rapidly from 8593 in 2016 to 90,566 in 2017 and gradually reduced to 23,481 in 2023. As previous studies showed that the two-dose EV-A71 vaccine can provide protection for about 2–5 y,^18,19,22,23 we assumed the protection period as 5 y in this study. The uptake of EV-A71 vaccine increased since 2017. Based on the annual survey of 0–5-y-old population in Shanghai Minhang, the estimated coverage rate of age-appropriate two-dose EV-A71 vaccination increased from 19.1% in 2017 to 53.5% in 2019, and reached 69.2% in 2021, then maintained over 65% but slightly dropped to 62.3% in 2023 (Figure 3). The coverage rate among children aged 0–5 y was significantly higher in Shanghai household registered children than in migrant children (67.8% vs. 52.6%, χ² = 4.101, p = .043). The lower vaccination coverage among migrant children is primarily due to their frequent mobility, relatively low family income and low level of parental education.

Figure 3.

The total number of EV71 vaccine doses administered and the coverage rates of EV71 vaccination among children aged 0–5 y in Shanghai Minhang district, 2016–2023.

Between January 2017 and December 2023, 3599 children with HFMD/HA were prospectively enrolled for the virological surveillance of enterovirus serotypes, including 2615 HFMD (1848 mild outpatients, 734 nonsevere inpatients, 44 severe inpatients) and 984 herpangina (902 mild outpatients, 61 nonsevere inpatients, and 1 severe inpatients). A total of 1174 cases were excluded, including 304 older than 6 y old or less than 6 months old, 29 residing outside Shanghai, 295 with unknown EV-A71 vaccination status and 546 pan-enterovirus negative. Finally, 2425 HFMD/HA cases were included for VE analysis, of whom, 32 (1.3%) were severe inpatients, 534 (22.0%) were nonsevere inpatients and 1859 (76.7%) were mild outpatients. EV-A71 was tested positive in 25 (78.1%) severe patients, 75 (14.0%) nonsevere inpatients, and 26 (1.4%) mild outpatients, respectively; 868 (35.8%) had two-dose EV-A71 vaccination, 129 (5.3%) had 1-dose EV-A71 vaccination, and 1428 (58.9%) were unvaccinated (Figure 4).

Figure 4.

Flow chart of subject enrollment in the test-negative design case-control study for the estimates of EV71 vaccine effectiveness during 2017–2023, in Shanghai Minhang district.

There was no significant difference in vaccination status and EV-A71 infection by gender and age groups (Table 2). Compared to unvaccinated cases, vaccinated cases were significantly less likely to contract HFMD/herpangina (68.0% vs. 85.7%, p < .001) prior to COVID-19 (2017–2019); vaccinated cases were significantly less likely to need hospitalization (17.7% vs. 25.1%, p < .001), develop severe complications (0.6% vs. 1.8%, p < .001), and contract EV-A71 infection (0.9% vs. 8.2%, p < .001). There was no statistically significant association between EV-A71 vaccination status and CV-A6, CV-A16, or CV-A10 related HFMD/herpangina.

Table 2.

Characteristics of vaccinated and unvaccinated pathogen-confirmed HFMD/herpangina children during 2017–2023.

Characteristics	Vaccinated group (N = 997) N (%)	Unvaccinated group (N = 1428) N (%)	P-value
Gender
Male	612 (61.4)	897 (62.8)	.475
Female	385 (38.6)	531 (37.2)
Age groups
6–35 months	480 (48.1)	694 (48.6)	.825
36–59 months	517 (51.9)	734 (51.4)
Consultation time
2017–2019 (before COVID-19)	678 (68.0)	1224 (85.7)	<.001
2020–2023 (after COVID-19)	319 (32.0)	204 (14.3)
Clinical diagnosis
HFMD	833 (83.6)	1077 (75.4)	<.001
Herpangina	164 (16.4)	351 (24.6)
Disease severity
Mild outpatients	815 (81.7)	1044 (73.1)
Nonsevere inpatients	179 (17.7)	355 (25.1)	<.001
Severe inpatients	3 (0.6)	29 (1.8)
EV-A71
+	9 (0.9)	117 (8.2)	<.001
−	988 (99.1)	1311 (91.8)
CV-A16a
+	163 (16.5)	187 (14.3)	.144
−	825 (83.5)	1122 (85.7)
CV-A6a
+	628 (44.1)	360 (41.2)	.178
−	796 (55.9)	513 (58.8)
CV-A10a
+	52 (5.3)	72 (5.5)	.479
−	936 (94.7)	1237 (94.5)

Note: EV-A71; enterovirus A71; CV-A16: coxsackievirus A16; CV-A6: coxsackievirus A6; CV-A10: coxsackievirus A10.^a126 EV-A71 positive cases were excluded.

The overall VE against EV-A71-associated HFMD/herpangina was calculated to be 89.9% (95% CI: 74.8–96.0) for two-dose vaccination and 67.9% (95% CI: 2.6–89.4) for one-dose vaccination (Table 3). The VE for two-dose vaccination was estimated to be 88.3% (95% CI: 61.9–96.4) among children aged 6–35 months and 91.6% (95% CI: 63.7–98.1) in the 36–59-month group. VE for two-dose vaccination was 89.3% (95% CI: 72.6–95.8) against HFMD and 100.0% against herpangina. VE for two-dose vaccination was 92.2% (95% CI: 41.8–98.9) against mild disease, 88.1% (95% CI: 66.8–95.8) against nonsevere hospitalization and 100% against severe disease, respectively. The collinearity analysis demonstrated that all the covariates (gender, age groups, consultation time, clinical diagnosis, and disease severity) in the logistic regression model exhibited tolerance exceeding 0.1 (0.845–0.995) and VIF below 10 (1.005–1.183), confirming the absence of multicollinearity. The number of events per variable is 485 (exceeding 10) in the logistic regression model, which indicates that the estimation of VE is robust.

Table 3.

Estimates of EV-A71 vaccine effectiveness against EV-A71 associated HFMD/herpangina by different patients’ charactistics during 2017–2023.

Analysis	Two-dose vaccination				One-dose vaccination
	Cases No./Total %	Controls No./Total %		VE (95% CI)	Cases No./Total %	Controls No./Total %	VE (95% CI)
Overall VEa,c,e	9/126 (7.1)		868/2175 (39.9)	89.9 (74.5–96.0)	4/121 (3.3)	124/1431 (8.7)	67.9 (2.6 –89.4)
By age groupsa,c,f
6–35 months	3/61 (4.6)		390/1024 (38.1)	88.3 (61.9–96.4)	3/61 (4.6)	86/720 (11.9)	68.5 (−20.6–91.8)
36–59 months	2/61 (3.2)		478/1151 (41.5)	91.6 (63.7–98.1)	1/60 (1.6)	38/711 (5.3)	69.2 (−140.8–96.1)
By consultation timea,c,j
2017–2019(Before COVID-19)	5/122 (4.1)		574/1677 (34.2)	89.9 (74.5–96.0)	4/121 (3.3)	99/1202 (8.2)	67.9 (2.6–89.4)
2020–2023(After COVID-19)	0/0 (0.0)		294/498 (59.0)	–	0/0 (0.0)	25/229 (10.9)	–
By clinical diagnosisa,c,h
HFMD	5/113 (4.4)		724/1689 (42.9)	89.3 (72.7–95.8)	4/112 (3.6)	104/1069 (9.7)	65.4 (−8.0–88.9)
HA	0/9 (0.0)		144/486 (29.6)	100.0	0/9 (0.0)	20/362 (5.5)	100.0
By clinical severitya,c,g
Mild outpatients	1/25 (4.0)		726/1743 (41.7)	92.2 (41.8–98.9)	1/25 (4.0)	90/1107 (8.1)	47.9 (−291.4–93.1)
Non-severe inpatients	4/73(5.5)		141/436 (32.3)	88.1 (66.8–95.8)	2/71 (2.7)	33/318 (10.4)	74.1 (−10.9–94.0)
Severe inpatients	0/24 (0.0)		1/6 (16.7)	100.0	1/25 (4.0)	1/6 (16.7)	75.0 (−426.2– 98.8)
By time since last vaccination
≥28 d post-vaccinationa,c,e	5/122 (4.1)		868/2175 (39.9)	89.9 (74.5–96.0)	4/121 (3.3)	124/1431 (8.7)	67.9 (2.6–89.4)
<28 d post-vaccinationb,c,e	5/121(4.1)		903/2192 (41.2)	90.2 (75.4–96.1)	5/121 (4.1)	107/1396 (7.7)	40.6 (−64.3–78.5)
By control subgroupa,d,e
Controls who were CV-A16 positive	5/122 (4.1)		147/333 (44.1)	91.5 (76.3, 96.9)	4/121 (3.3)	17/203 (8.4)	55.5 (−90.5, 89.6)
Controls who were CV-A6 positve	5/122 (4.1)		551/1347 (40.9)	90.1 (73.9, 96.2)	4/121 (3.3)	78/874 (8.9)	65.6 (−8.4, 89.1)
Controls who were CV-A10 positive	5/122 (4.1)		47/119(39.5)	92.4 (78.0, 97.4)	4/121 (3.3)	5/77 (6.5)	66.4 (−84.2, 93.9)
Controls who were EV positive	5/122 (4.1)		128/381 (35.6)	90.1 (74.4, 96.2)	4/121 (3.3)	22/275 (8.0)	71.9 (1.3, 92.0)

^aVE analysis against EV-A71 infection considering 28-d vaccination lags; ^bVE analysis against EV-A71 infection including all vaccinated <28 d before onset; ^ccontrols of TND study were positive for an enterovirus other than EV-A71; ^dcontrols of TND study were positive for CV-A16, CV-A6, CV-A10, or an enterovirus other than EV-A71, CV-A16, CV-A6, and CV-A10; ^eadjusted for gender, age groups, consulting time, clinical diagnosis, and clinical severity; ^fadjusted for gender, consulting time, clinical diagnosis, and clinical severity; ^gadjusted for gender, age groups, consulting time, and clinical dignosis; ^hadjusted for gender, age groups, consulting time, and clinical severity; ⁱadjusted for gender, age groups, clinical diagnosis, and clinical severity.

The VE estimates were similar in the sensitivity analysis for two-dose EV-A71 vaccination within 28 d and post-28 d. Also, we found that estimated VEs did not change significantly with different control group criteria. Differences in adjusted VEs varied by less than 3% points from the primary analysis control group for each of the alternative negative control group definitions (Table 3).

Discussion

The national surveillance data from China CDC reported a remarkable decline of EV-A71-HFMD (28.3%), severity rate (62.2%), and mortality rate (83.8%) of HFMD patients by vaccination program (2017–2019) compared to 2013–2016.²⁴ Our study revealed much more reduction in the incidence rate (56.7%), case-severity rate (95.2%), and mortality rate (100%) in Shanghai during the postvaccination period (2017–2023) compared to prevaccination period (2009–2016). The similar decreasing trend was also reported in other regions, such as Fujian²⁵ (47.6% reduction in 2018–2021 vs. 2014–2017) and Chengdu²⁶ (60.0% reduction in 2017–2018 vs. 2011–2016), while notably higher than that reported in Zhejiang²⁷ (29.0% reduction in 2017–2020 vs. 2010–2016). The observed variations may be explained by the regional variations in vaccination coverage rates among young children. Two-dose EV-A71 vaccination coverage rate was around 60.0% in Shanghai from 2019 to 2023, 54.8% in Fujian from 2018 to 2021 and 54.3% in Chengdu in 2018. By contrast, the vaccination coverage rate in Zhejiang in 2019 was only 17%.^25–27 These findings suggested enhanced protective effectiveness against HFMD with an increasing uptake rate of EV-A71 vaccine. Of particular note, severe and fatal cases were almost prevented in Shanghai in the era of EV-A71 vaccine. However, in low-income region, such as Guizhou, locating the southwest region of China, where EV-A71 remained endemic and associated fatal and severe cases were reported during 2019–2023.²⁸ This different situation is more likely to be related to low vaccination coverage.

Although the 0–5-y-old age group consistently accounted for the highest proportion throughout the 15-y study period, we observed a decrease of 13.2% in the proportion of this age group, while the proportion of the 6–10-y-old age group increased by 105.7% and that of school-age children increased two-fold since the implementation of EV-A71 vaccination. The national surveillance data showed that a 3-y EV-A71 vaccination program led to a rise in the median age of HFMD infection by 0.57 y,²⁴ which paralleled a 0.56-y increase observed during a consecutive 7-y vaccination program in Shanghai. The gradual expansion of the susceptible age group to school-age children may reflect two facts. First, the herd immunity barrier was enhanced among children aged 0–5 y old through EV-A71 vaccination or natural infection, as a consequence, reducing the risk of EV-A71 infection among young children; however, immunity gap makes school-age children more susceptible to HFMD. A 5-y cohort study including 211 participants conducted in Jiangsu Province showed that the vaccinated group had significantly higher serum antibody levels than the unvaccinated group (GMT: 141.42 vs. 71.83; seropositivity rate: 94.34% vs. 71.43%),²⁹ suggesting that EV-A71 vaccination can significantly induce enhanced levels of IgG antibodies for vaccinees. Furthermore, a routine serological surveillance of EV-A71 antibodies among 100 healthy children in the Minhang district, sponsored by the Shanghai CDC in 2024 (nonpeer-reviewed unpublished internal monitoring data), revealed a higher seropositivity rate in children aged 1–5 y than in those aged 6–12 y (62.5% vs. 55.0%), with an average two-dose EV-A71 vaccination coverage of 51.42% among children aged 0–5 y in the same year. Combining the results of the aforementioned cohort study and our surveillance data, we speculate that the EV-A71 vaccination, in the context of high coverage, possibly builds an immune protective barrier among susceptible young children, then consequently leading to a shift in susceptibility from younger, vaccine-eligible children to older unvaccinated children. As this hypothesis is derived from indirect unpublished small-size serological data, we need to carry out the well-designed prospective cohort or case-control studies to further verify the validity of this hypothesis, based on the raw serological data and epidemiological investigation findings. Second, school-age children who may not contract CV-A6 and CV-A10 previously; thus, they were more susceptible to the dominantly circulating CV-A6 and CV-A10 serotypes. A recent study in the USA reported a CV-A6-associated HFMD outbreak involving 138 college students.³⁰ The 15-y virological surveillance data of HFMD revealed that the predominant pathogens shifted from EV-A71/CV-A16 (2009–2014) to CV-A6/CV-A16/EV-A71 (2015–2017), and almost replaced by CV-A6/CV-A16 after 2018, and EV-A71 serotype was undetectable in the sampled cases since 2019 in Shanghai. Therefore, the shift of susceptible age and the replacement of dominant serotype are worthy of attention and close monitoring. The application of multivalent vaccines and expanding vaccination eligibility to include school-age children will be helpful for further prevent and control HFMD.

Two-dose EV-A71 vaccination showed an overall 89.9% of robust protection effectiveness against EV-A71-associated HFMD/herpangina in children below 6 y of age over 7-y vaccination program, regardless of the disease severity. The estimated VE was lower than the VE in Phase III clinical trials against EV-A71-associated diseases in 2014 (94.8%),¹⁰ but higher than the real-world VE estimates in Henan against HFMD inpatients during 2017–2018 (85.4%)¹⁷ and in Beijing against non-severe HFMD outpatients in 2017 (83.7%).¹⁸ These difference could be explained by variation factors, such as enrolled subjects and disease severity, observation periods, and vaccination coverage rates. Also, we observed the overall VE was lower for a single dose of EV-A71 vaccination (67.9%) than that for two series vaccination (89.9%), suggesting that one-dose can offer partial protection against EV-A71 infection. This finding is consistent with that reported by previous studies.^8,17–19 Thus, the real-world study also supports that two-dose EV-A71 vaccination is optimal and necessary for robust immune protection. Besides, we found no significant difference in the effectiveness of EV-A71 vaccine given at the age of 36–59 months and the age of 6–35 months, despite that the VE was slightly higher for older children. The increasing point estimates of VE with increasing age were consistent with the phenomenon that older children showed a higher geometric mean titer (GMT) than younger children in the Phase III trials.⁹ Preschool children aged 3 y and older have possibly contracted inapparent EV-A71 infection prior to vaccination and have preexisting EV-A71 antibodies.^29,31 EV-A71 vaccination can booster preexisting immunity and thus offers better protection. When stratifying the analysis by disease severity, VE was slightly higher against mild outpatients than against non-severe inpatients, probably associated with the much higher two-dose EV-A71 vaccination rates among the mild outpatients. We also noticed that vaccinated cases were significantly less likely to need hospitalization and develop severe complications, strongly suggesting that vaccine-induced immune protection mitigates disease severity.³² Of particular note, we didn’t confirm cross-protection effectiveness for EV-A71 vaccines against CV-A6, CV-A16, or CV-A10.

There were several strengths in this study. First, this study spanned a 15-y observational period in Shanghai and integrated epidemiological, virological, and vaccination surveillance data to evalute EV-A71 vaccination impact on HFMD. Second, this study provided the comprehensive real-world evidence on assessing VE of EV-A71 vaccination across varying clinical severities of HFMD/herpangina over seven epidemic seasons. Third, all enrolled children with HFMD/herpangina included in this test-negative case-control study for VE estimates were virological diagnosed cases, thus improving the results robustness. Meanwhile, several limitations should be considered when interpreting and generalizing our research findings. First, although we verified EV71-vaccination history through both electronic records and guardians’ memory, guardians’ recall bias may have led to misclassification of vaccination status.³³ On the other hand, the migrant population count obtained via SMSIIP is an imprecise estimate. A minor migrant children were not included in SMSIIP if they didn’t receive vaccination while temporarily residing in Minhang district, or if their immunization records had not yet been created within 3 months after the first vaccination. Second, enrolled children might ever have EV-A71 subclinical or asymptomatic infection, which could lead to overestimate VE among children aged 36–59 months. A prospective birth cohort study showed that the cumulative incidence rate was 15% by 36 months of age, and 29% of EV-A71 infections were asymptomatic in young children.³¹ However, it is difficult to differentiate the immune protection elected by subclinical infection from vaccination. Theoretically, vaccination can booster immunity induced by natural infection for a vaccinated child who had inapparent EV-A71 infection prior to EV-A71 vaccination. Third, we didn’t do EV-A71 sub-genotyping and virus neutralization test in this study. Based on the national surveillance data from 2009 to 2018, C4a accounted for 88.78% of subgenotype and C4b, C1, C2, C5, B0–B5 also co-circulated.³⁴ In a phase IV clinical study involving 140 participants vaccinated with EV-A71 vaccines, C4 EV-A71 vaccines were demonstrated to provide global protection to infants and children against HFMD caused by different EV-A71 subgenotypes (A, B0-B4, C1, C2, C4, and C5 viruses).³⁵ Therefore, it is reasonable to infer that the current EV-A71 vaccine provides neutralizing protection against the predominant EV-A71 genotypes and sub-genotypes circulating in Shanghai and across the country. Finally, we acknowledge that the implementation of non-pharmaceutical interventions (NPIs) during the COVID-19 pandemic (2020–2022) undoubtedly had a significant impact on reducing HFMD incidence.³⁶ Although HFMD is a notifiable disease in China, post-pandemic management of HFMD in schools/kindergartens has been further strengthened, which likely minimized underreporting of severe cases, we cannot rule out the possibility of undereporting of mild cases. This potential underreporting could have influenced the observed HFMD incidence trends, particularly during the periods of the COVID-19 pandemic when health-care resources were strained. However, when the incidence of HFMD rebounded in 2023, no EV-A71 serotyped were detected in monitoring cases. This monitoring data indirectly illustrates that the EV-A71 vaccine has a good protective effect against EV71-related diseases in real-world applications in Shanghai.

In conclusion, our findings project the long-term impact of the EV-A71 vaccination on the HFMD epidemic pattern in a high-density urban area in Shanghai with around 50% of coverage rate among young children. A dramatic decline in EV-A71-associated HFMD, particularly severe cases, was observed in Shanghai after the introduction of the EV-A71 vaccine. Two-dose EV-71 vaccination showed a good real-world protection effectiveness against EV-A71-associated diseases in age-appropriate immunized young children. Given the wide use of EV-A71 vaccines in China, it warrants continuous monitoring the epidemiological trend of HFMD and the vaccine-driven shift of predominant serotypes, as well as its impact on disease severity and susceptible population. In the future, multivalent vaccines development is promising to further control the epidemics of HFMD.

Biographies

Dongli Xu, MD, Associate Chief Physician, currently serves as the Party Secretary of Minhang District Center for Disease Control and Prevention (Minhang Health Supervision Institute). She has extensive experience in various fields, such as health education, infectious disease prevention and control, AIDS prevention and control, and chronic disease prevention and control. She played a pivotal role in talent development and health-care resource allocation and management for district-level public health systems.

Mei Zeng, currently serving as the Director of the Department of Infectious Diseases at Children’ Hospital of Fudan University, professor of Fudan University Biosecurity Research Institute, is chair of Pediatric Subgroup of China Society of Infectious Diseases of Chinese Medical Association, vice chair of Pediatric Immunization Committee of Chinese Medical Association, and an expert member of WHO EML Antibiotic Working Group. Her major areas of interest focus on the epidemiological studies of childhood vaccine-preventable infectious diseases, emerging and re-emerging infectious diseases, antibiotic stewardship, and vaccine-related issues. She published more than 120 academic papers as the first author and the corresponding author.

Funding Statement

This work was supported by the Natural Science Research Project in Minhang district, Shanghai [No. 2024MHZ033], Talent Development Program for Outstanding Public Health Professionals in Minhang district, Shanghai [No. 20240607], Construction and Cultivation of Brand Departments in Minhang District [No. MGWKS01], Infectious Disease Surveillance and Prevention Work in Shanghai Municipality (2009–2023), National Major Infectious Disease Surveillance Work (2009–2023) and 2023 Shanghai Oriental Talent Program Youth Project.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Ethics approval statement

Sample testing and case investigation of HFMD and herpangina at the sentinel hospital were a routine surveillance in Shanghai. Informed consent was exempt from all investigated cases. This study was approved by the institutional ethical review board of Shanghai Minhang Center for Disease Control and Prevention (No. EC-2024-008).

Supplemental material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/21645515.2026.2656515