Hand, foot, and mouth disease (HFMD), which is mainly caused by coxsackievirus A16 (CVA16) or enterovirus A71 (EV-A71), poses a serious threat to children’s health. However, the long-term dynamics of the neutralizing Ab (NAb) response and ideal paired-serum sampling time for serological diagnosis of CVA16-infected HFMD patients were unclear. In this study, 336 CVA16 and 253 EV-A71 PCR-positive HFMD inpatients were enrolled and provided 452 and 495 sera, respectively, for NAb detection. Random-intercept modeling with B-spline was conducted to characterize NAb response kinetics. The NAb titer of CVA16 infection patients was estimated to increase from negative (2.1, 95% confidence interval [CI]: 1.4–3.3) on the day of onset to a peak of 304.8 (95% CI: 233.4–398.3) on day 21 and then remained >64 until 26 mo after onset. However, the NAb response level of EV-A71–infected HFMD patients was much higher than that of CVA16-infected HFMD patients throughout. The geometric mean titer was significantly higher in severe EV-A71–infected patients than in mild patients, with a 2.0-fold (95% CI: 1.4–3.2) increase. When a 4-fold rise in titer was used as the criterion for serological diagnosis of CVA16 and EV-A71 infection, acute-phase serum needs to be collected at 0–5 d, and the corresponding convalescent serum should be respectively collected at 17.4 (95% CI: 9.6–27.4) and 24.4 d (95% CI: 15.3–38.3) after onset, respectively. In conclusion, both CVA16 and EV-A71 infection induce a persistent humoral immune response but have different NAb response levels and paired-serum sampling times for serological diagnosis. Clinical severity can affect the anti–EV-A71 NAb response.

Hand, foot, and mouth disease (HFMD) is a common infectious disease posing a strong threat to children worldwide, mostly occurring in those <5 y of age (1), especially in China (2), with nearly 21 million patients reported between 2010 and 2019. The number of patients and deaths due to HFMD in China rank first in the word (13). Enterovirus A71 (EV-A71) and coxsackievirus A16 (CVA16) have been identified as the predominant causative pathogens (2), and they mainly lead to severe HFMD outbreaks in Asia (2, 4, 5). Although inactivated monovalent vaccines against EV-A71 have been licensed in mainland China since 2016, there is no cross-protective effect against CVA16 (6). In addition, the number of CVA16-related HFMD patients increased from 2017 to October 2019 (7, 8).

Neutralizing Abs (NAbs) play a crucial protective role in viral infection and are one of the most important indicators of humoral immunity. Therefore, understanding the dynamics of the NAb response in patients with HFMD caused by CVA16 and EV-A71 can shed light on the vaccination strategy for HFMD-related vaccines, such as the length of the protection period for patients with infection history and whether an additional vaccine is needed for these patients. However, seroepidemiological studies of enteroviruses related to HFMD have mainly focused on EV-A71; there are few studies on CVA16 (912), especially long-term observations, and most of them have been based on a cross-sectional design, focused on healthy children or used sera remaining from other projects. In addition, most longitudinal studies are based on sera taken from 1 wk to 1 mo postinfection or after vaccination and are often limited by small sample sizes (11, 12); thus, it is unclear how the immune response wanes/changes over a longer time.

A 4-fold increase in enterovirus-specific Ab concentrations in the recovery phase compared with that in the acute phase serves as laboratory confirmation of recent infection (13). However, no studies have systematically defined the acute and convalescence phases of enterovirus infection. Previous studies performed according to experience without much evidence have defined various periods, such as 0–3 d (11, 14, 15), 0–5 d (16), 0–7 d (17, 18) or 2–4 d (19) after onset, as the acute stage. Researchers have collected convalescent sera at 7–16 d (11, 19), 14–30 d (16), 4–6 wk (20) or 2 mo (14) after onset. In some patients, studies have indicated only the time interval of collecting acute and convalescent samples, such as 4–7 d (21) or 2 wk (13). Thus, this creates difficulty in research and clinical practice, resulting in a lack of clarity regarding when to collect blood specimens with productive specific Abs.

Thus, based on a prospective, hospital-based cohort study of HFMD patients, we describe the kinetics of the NAb response in EV-A71–infected patients and the ideal sampling time of paired sera for serological diagnosis of EV-A71 infection (22). In this study, we aimed to illustrate the kinetics of NAb levels in patients during the course of CVA16 infection for a long time, up to 1.5 y, to analyze the difference in the NAb response level and ideal sampling time of paired sera between CVA16- and EV-A71–infected patients.

A longitudinal cohort study of HFMD patients was conducted in Henan Children’s Hospital between 2017 and 2019 as previously described (22). In brief, HFMD inpatients were enrolled during the acute stage. Then, a subgroup of patients, excluding those who previously had conditions that may affect neurodevelopment, was invited to follow up at 2 wk, 3 mo, 6 mo, and 1.5 y after discharge. Demographic and clinical information was extracted from medical records using a standardized case record form. Severe HFMD patients were defined as follows: 1) having CNS complications, including meningitis, encephalitis, brainstem encephalitis, encephalomyelitis, acute flaccid paralysis, autonomic nervous system dysregulation, or subsequent severe cardiopulmonary complications (23); and 2) requiring admittance to the pediatric intensive care unit (PICU) during hospitalization.

Throat swabs and stool samples were collected for enterovirus infection diagnosis by RT-PCR as previously described (6, 22). Briefly, throat swabs were first tested using real-time RT-PCR for EV-A71 and/or CVA16. When negative, nested RT-PCR was used for further detection. When throat swabs tested negative, stool samples, if available, were detected by a commercial real-time RT-PCR kit (Mole Bioscience, Taizhou, China) targeting EV-A71, CVA16, and pan-enterovirus. CVA16 RNA single-positive HFMD patients with at least one serum sample collected were the subjects of this study. In addition, 264 EV-A71–infected patients were described in our previous study (22), and among them, 96% (253/264) were included for comparative analysis with CVA16 after excluding samples that were double positive for EV-A71 and CVA16.

The study protocol was reviewed and approved by the Ethics Committees of the Chinese Center for Disease Control and Prevention, Henan Children’s Hospital and School of Public Health, Fudan University (22).

Sera were collected at admission, during disease progression and at discharge, as well as at every follow-up visit after discharge (22) for NAb detection. We performed neutralization assays according to a standard protocol as previously described (22). In brief, serum was serially diluted 2-fold from 1:8 to 1:16,384 and mixed with an equal volume of 100 tissue culture infectious dose 50% of the CVA16/190 strain (GenBank accession number JF420555 [https://www.ncbi.nlm.nih.gov/nuccore/JF420555], genotype B1b) (24). After incubation at 37°C for 2 h, human rhabdomyosarcoma cells (1–2 × 105 cells/ml) were added to each well and incubated at 37°C for 4–7 d. All diluted samples were tested in duplicate, and the titration was double checked. NAb titers were defined as the reciprocal of the highest dilution capable of inhibiting 50% of the cytopathic effect and calculated by use of the Karber method (25).

NAb titers are presented as reciprocal titers throughout and were log2 transformed before analysis. Continuous variables were compared by the Wilcoxon rank sum test, and categorical variables were compared by using a χ2 test or Fisher exact test.

To calculate the seropositive percentage, median NAb titer, and geometric mean titer (GMT) of CVA16- and EV-A71–infected patients during the illness course, we excluded EV-A71–infected patients that were vaccinated against EV-A71. To model the Ab response kinetics from CVA16- and EV-A71–infected patients, we pooled serological data from patients with at least two sera and excluded patients who were still seronegative for 2 mo after illness onset to avoid the effect of false positives of PCR-diagnosed infection. A random-intercept model with B-spline (generalized linear model [GLM]) was established to characterize the anti-CVA16 and anti–EV-A71 NAb response over time in HFMD patients after onset. A NAb titer ≥16 was considered seropositive (26), and sensitivity analyses were also performed using cutoffs of 8 and 32.

Three parametric distributions (Weibull, γ, and log-normal) with maximum likelihood estimation were used to estimate the distribution of the disease course and healthcare-seeking behavior of HFMD patients. Typical symptom recession was considered the acute phase, and the corresponding covalence phase was fit with the three aforementioned distributions considering data censoring. All analyses were performed in R and SAS. The GLM selection and the best fit distribution selection were based on the Akaike information criterion. The acute phase was defined as the average time period from illness onset to the subsidence of typical symptoms. The corresponding convalescent period was defined as the average time at seroconversion, 4-fold increase, and 8-fold increase in the first NAb appearance compared with each day in the acute phase. For the missing data in the acute phase, we conducted an average imputation with the GLM mentioned above and shortest distance decision filling. Two imputation methods were used for the sensitivity analyses as described in the Supplemental Tables I and II. To minimize the effect of unbalanced follow-up, the parametric estimation considered the interval and right censoring of data, and 2 mo was treated as the final event time. The bootstrap method was used to estimate the 95% CI of the convalescent period. A p values of <0.05 was considered statistically significant.

The funders had no role in the study design, data collection, data analysis, data interpretation, or writing of the manuscript. The corresponding author had full access to all the data and had the final responsibility for the decision to submit for publication.

Overall, 343 (19%) of 1840 HFMD inpatients tested positive for CVA16 by RT-PCR and were enrolled in our study (Fig. 1). For CVA16-infected patients, 98% (336/343) provided 452 sera. Of these, 371 sera were collected from 335 patients during hospitalization, and 81 sera were collected from 39 patients during the follow-up period. Of these, 36 (92%), 13 (33%), 19 (49%), and 13 (33%) patients participated in the 2-wk, 3-mo, 6-mo, and 1.5-y follow-ups, respectively (Fig. 1, Supplemental Fig. 1). A total of 495 sera from 253 EV-A71 RNA-positive HFMD inpatients were included for comparative analysis.

FIGURE 1.

Flowchart of participant enrollment and follow-up. #1Indicates that 302 patients provided one serum sample, 31 patients provided two serum samples, 1 patient provided three serum samples, and 1 patient provided four serum samples during hospitalization. #2Indicates that 73 patients without complications met the maximum enrollment number; 52 met neurodevelopmental exclusion criteria, including 19 patients who were premature, 31 patients who had a history of PICU admission/ventilation, 8 patients who had any prior chronic respiratory, cardiac, or other illnesses (e.g., congenital hypothyroidism, congenital epilepsy, and asthma), 6 patients who had prior developmental or neurodevelopmental delays, and 2 patients who had prior learning disabilities or neurologic regression. #3Indicates that 37 patients were recovering, 35 patients were unable to be contacted for follow-up, 44 patients declared that they had no time to participate in follow-up, 32 patients thought the project lasted too long, 2 patients were scared of the hospital or blood drawing, and 15 patients refused to participate in the follow-up because they were too far away. *Indicates that these subjects participated in the 2-wk follow-up but refused serum collection.

FIGURE 1.

Flowchart of participant enrollment and follow-up. #1Indicates that 302 patients provided one serum sample, 31 patients provided two serum samples, 1 patient provided three serum samples, and 1 patient provided four serum samples during hospitalization. #2Indicates that 73 patients without complications met the maximum enrollment number; 52 met neurodevelopmental exclusion criteria, including 19 patients who were premature, 31 patients who had a history of PICU admission/ventilation, 8 patients who had any prior chronic respiratory, cardiac, or other illnesses (e.g., congenital hypothyroidism, congenital epilepsy, and asthma), 6 patients who had prior developmental or neurodevelopmental delays, and 2 patients who had prior learning disabilities or neurologic regression. #3Indicates that 37 patients were recovering, 35 patients were unable to be contacted for follow-up, 44 patients declared that they had no time to participate in follow-up, 32 patients thought the project lasted too long, 2 patients were scared of the hospital or blood drawing, and 15 patients refused to participate in the follow-up because they were too far away. *Indicates that these subjects participated in the 2-wk follow-up but refused serum collection.

Close modal

The ages of CVA16-infected patients ranged from 2 mo to 12 y, and 54% (182/336) of patients were <2 y old. Male patients predominated (223 males/113 females). Of the patients, 2% of patients required i.v. Ig (IVIG) treatment. The median interval between illness onset and admission was 2 d (interquartile range [IQR], 1–3 d), and the median length of hospital stay (LOS) was 4 d (IQR, 4–5 d). Few patients experienced severe illness (5%, 16/336); among them, 94% (15/16) had CNS complications, and 25% (4/16) were admitted to the PICU (Table I). Compared with CVA16-infected patients who were not invited or refused to participate in the follow-up, there was no significant difference in demographic characteristics.

Table I.

Baseline characteristics of CVA16-infected patients who provided sera during hospitalization and follow-up

CharacteristicsCVA16 (n = 336)Provided Sera during Hospitalization (n = 335)Participated in Follow-Up and Provided Sera (n = 39)
Male, n (%) 223 (66) 222 (66) 28 (72) 
Female, n (%) 113 (34) 113 (34) 11 (28) 
Age, y, n (%)    
 <2 182 (54) 182 (54) 20 (51) 
 2–12 154 (46) 153 (46) 19 (49) 
With EV-A71 vaccination, n (%) 51 (15) 51 (15) 8 (21) 
Clinical course    
Interval (days) between illness onset and admission, median (IQR) 2 (1, 3) 2 (1, 3) 2 (1, 3) 
Median (IQR) LOS, days 4 (4, 5) 4 (4, 5) 4 (4, 5) 
Requirement of IVIG treatment, n (%) 6 (2) 6 (2) 3 (8) 
Clinical severity (severe), n (%) 16 (5) 16 (5) 4 (10) 
 CNS complications 15 (94) 15 (94) 4 (100) 
 PICU admission during hospitalization 4 (25) 4 (25) 1 (25) 
CharacteristicsCVA16 (n = 336)Provided Sera during Hospitalization (n = 335)Participated in Follow-Up and Provided Sera (n = 39)
Male, n (%) 223 (66) 222 (66) 28 (72) 
Female, n (%) 113 (34) 113 (34) 11 (28) 
Age, y, n (%)    
 <2 182 (54) 182 (54) 20 (51) 
 2–12 154 (46) 153 (46) 19 (49) 
With EV-A71 vaccination, n (%) 51 (15) 51 (15) 8 (21) 
Clinical course    
Interval (days) between illness onset and admission, median (IQR) 2 (1, 3) 2 (1, 3) 2 (1, 3) 
Median (IQR) LOS, days 4 (4, 5) 4 (4, 5) 4 (4, 5) 
Requirement of IVIG treatment, n (%) 6 (2) 6 (2) 3 (8) 
Clinical severity (severe), n (%) 16 (5) 16 (5) 4 (10) 
 CNS complications 15 (94) 15 (94) 4 (100) 
 PICU admission during hospitalization 4 (25) 4 (25) 1 (25) 

In contrast to CVA16-infected patients, more than half of EV-A71–infected patients were >2 y old (EV-A71 of 57% versus CVA16 of 46%, χ2 test, p = 0.010), and there was a higher proportion of severe cases (EV-A71 40% versus CVA16 5%, χ2 test, p < 0.001) that required IVIG treatment (EV-A71 26% versus CVA16 2%, Fisher exact test, p < 0.001). Among severe EV-A71–infected patients, all patients had CNS complications, and 39% (40/102) of patients were admitted to the PICU.

Most CVA16-infected patients (95%, 61/64) had symptoms of rash and fever, and other patients had only rash symptoms. The average durations of fever and rash were 2.3 d (95% CI: 0.4–11.8 d) and 4.7 d (95% CI: 1.9–9.4 d), respectively. The average duration from symptom onset to symptom disappearance was 5.1 d (95% CI: 2.4–11.0 d). However, all of the durations mentioned above were significantly (1–3 d) shorter than those in EV-A71–infected patients (Table II).

Table II.

The average time interval (in days) and 95% CI of the natural history and medical behavior of CVA16- and EV-A71–infected patients

Time IntervalCVA16 (n = 336)EV-A71 (n = 253)ap Valueb
Natural history of disease    
 Duration of the fever/rash until disappearance of all symptoms 5.1 (2.4–11.0) 7.0 (3.0–16.7) <0.001 
 Duration of fever 2.3 (0.4–11.8) 5.0 (1.3–19.8) <0.001 
 Duration of rash 4.7 (1.9–9.4) 5.7 (1.5–11.1) 0.016 
 Interval between rash and fever 0.6 (0.3–1.5) 0.7 (0.2–2.4) 0.663 
Medical behavior    
 Duration from onset to hospital admission 1.6 (0.4–6.4) 2.8 (0.5–6.9) <0.001 
 Hospitalization duration 4.9 (2.9–8.4) 7.0 (3.0–16.7) <0.001 
 Duration from onset to discharge 6.8 (3.8–12.3) 10.1 (4.9–20.6) <0.001 
Time IntervalCVA16 (n = 336)EV-A71 (n = 253)ap Valueb
Natural history of disease    
 Duration of the fever/rash until disappearance of all symptoms 5.1 (2.4–11.0) 7.0 (3.0–16.7) <0.001 
 Duration of fever 2.3 (0.4–11.8) 5.0 (1.3–19.8) <0.001 
 Duration of rash 4.7 (1.9–9.4) 5.7 (1.5–11.1) 0.016 
 Interval between rash and fever 0.6 (0.3–1.5) 0.7 (0.2–2.4) 0.663 
Medical behavior    
 Duration from onset to hospital admission 1.6 (0.4–6.4) 2.8 (0.5–6.9) <0.001 
 Hospitalization duration 4.9 (2.9–8.4) 7.0 (3.0–16.7) <0.001 
 Duration from onset to discharge 6.8 (3.8–12.3) 10.1 (4.9–20.6) <0.001 
a

The natural history and medical behavior of EV-A71–infected patients were similar to those described in our previous study (22). Nuances in numerical value due to 11 patients were both positive for EV-A71 and CVA16 according to PCR and were excluded.

b

A Wilcoxon rank sum test was used to test the difference between two serotypes.

Regarding medical behavior, the average time from onset to hospital admission and discharge was 1.6 d (95% CI: 0.4–6.4 d) and 6.8 d (95% CI: 3.8–12.3 d), respectively, and the corresponding average LOS was 4.9 d (95% CI: 2.9–8.4 d), which was also significantly shorter than that of EV-A71 patients (Wilcoxon rank sum test, p < 0.001, Table II).

The CVA16 seropositive percentage remained <25% within 3 d of illness onset. This percentage significantly increased to almost 50% (49.0%, 95% CI: 35.3–62.9%) on day 4. Thereafter, all patients became seropositive between 11 d and 6 mo, and a slight decline was observed after 6 mo (92.3%, 95% CI: 64.0–99.8%) (Fig. 2A). Different cutoffs (1:8, 1:16, and 1:32) had little impact on prevalence estimates except for the difference between the cutoffs 8 and 32 at day 2. The anti-CVA16 NAb titer converted from negative (median, 6.0; IQR, 6.0–6.0; GMT, 6.5) on day 1 to positive (median, 11.0; IQR, 6.0–107.9; GMT, 26.1) on day 4 (Wilcoxon signed-rank test, p < 0.001). The NAb level rapidly increased and reached its peak 11–20 d after onset (median, 304.4; IQR, 139.6–512.0; GMT, 272.7). Although the NAb level decreased slightly after the peak titer, the median and GMT of the anti-CVA16 NAb titer showed no significant difference between each time point after 20 d (Fig. 2B).

FIGURE 2.

Seropositive percentage and NAb response of CVA16-infected patients over the illness course in contrast to those of EV-A71–infected patients. (A) Bars indicate seropositive percentages and their 95% confidence intervals. (B) NAb response. Black dotted lines indicate the GMT of NAb. The number of serum samples is listed below each time point. The χ2 test and Wilcoxon rank sum test were used to test the difference in the positive percentage and median of titer level, respectively, between CVA16 and EV-A71. Additionally, the McNemar test and Wilcoxon signed-rank test were used to test those differences between adjacent time intervals for each serotype. The gray dashed line represents the positive threshold for EV-A71 patients and CVA16 patients. *p < 0.05, **p < 0.01, ***p < 0.001. An m in the x-axis labels represents months. Note that 468 sera from 243 EV-A71–infected patients and 452 sera from 336 CVA16-infected patients were used in this analysis.

FIGURE 2.

Seropositive percentage and NAb response of CVA16-infected patients over the illness course in contrast to those of EV-A71–infected patients. (A) Bars indicate seropositive percentages and their 95% confidence intervals. (B) NAb response. Black dotted lines indicate the GMT of NAb. The number of serum samples is listed below each time point. The χ2 test and Wilcoxon rank sum test were used to test the difference in the positive percentage and median of titer level, respectively, between CVA16 and EV-A71. Additionally, the McNemar test and Wilcoxon signed-rank test were used to test those differences between adjacent time intervals for each serotype. The gray dashed line represents the positive threshold for EV-A71 patients and CVA16 patients. *p < 0.05, **p < 0.01, ***p < 0.001. An m in the x-axis labels represents months. Note that 468 sera from 243 EV-A71–infected patients and 452 sera from 336 CVA16-infected patients were used in this analysis.

Close modal

Compared with the anti–EV-A71 NAb dynamic (22), the CVA16 seropositive percentage was significantly lower from day 1 to day 10 but similar at 11 d after onset (Fig. 2A). However, the median and GMT of the anti-CVA16 NAb titer were always significantly lower than those of the anti-EV-A71 NAb titer at each time point during the whole study period (Fig. 2B).

The model prediction indicated that CVA16-infected patients did not produce an effective NAb response (titer, 2.1; 95% CI: 1.4–3.3) on the day of illness onset and that they became positive by day 3 (NAb titer, 28.3; 95% CI: 21.1–37.9). NAb levels increased and peaked (304.8; 95% CI: 233.4–398.3) on day 21 (Fig. 3B) and then remained >64 until 26 mo after illness onset (Fig. 3A). In addition, anti-CVA16 NAb response kinetics were coincident in general except that the peak titer was a slightly lower (170 versus 305) when patients with only one serum sample were added to the model.

FIGURE 3.

Prediction of anti-CVA16 and anti–EV-A71 NAb responses. (A) NAb responses over the whole time period. (B) NAb responses within 40 d after illness onset. The curves and gray ribbons represent the average NAb titer and 95% CI. The horizontal line indicates the positivity threshold. The black diamond points in the top of the panel indicate the peak time and maximum Ab titer for EV-A71 and CVA16. In the bottom of the panel, the lowercase letters (a–d) represent the average admission time (1.6 d), dates on which the NAb titers achieved the positivity threshold (2.5 d), average discharge time (6.8 d), and peak value of the Ab titer (21 d) for CVA16, respectively. The capital letters (A, C, and D) represent the average admission time (2.8 d), the average discharge time (7.0 d), and the peak value of the Ab titer (13 d) for EV-A71, respectively. Note that the model of anti-CVA16 NAb dynamics included 178 sera from 64 patients who provided serial samples throughout, and the model of anti–EV-A71 NAb dynamics included 310 sera from 92 patients who provided serial samples throughout.

FIGURE 3.

Prediction of anti-CVA16 and anti–EV-A71 NAb responses. (A) NAb responses over the whole time period. (B) NAb responses within 40 d after illness onset. The curves and gray ribbons represent the average NAb titer and 95% CI. The horizontal line indicates the positivity threshold. The black diamond points in the top of the panel indicate the peak time and maximum Ab titer for EV-A71 and CVA16. In the bottom of the panel, the lowercase letters (a–d) represent the average admission time (1.6 d), dates on which the NAb titers achieved the positivity threshold (2.5 d), average discharge time (6.8 d), and peak value of the Ab titer (21 d) for CVA16, respectively. The capital letters (A, C, and D) represent the average admission time (2.8 d), the average discharge time (7.0 d), and the peak value of the Ab titer (13 d) for EV-A71, respectively. Note that the model of anti-CVA16 NAb dynamics included 178 sera from 64 patients who provided serial samples throughout, and the model of anti–EV-A71 NAb dynamics included 310 sera from 92 patients who provided serial samples throughout.

Close modal

Compared with the anti–EV-A71 NAb response, the anti-CVA16 NAb response grew at a relatively slow rate and remained at a lower level over time (Fig. 3). After adjusting for age and clinical severity, the NAb response level for EV-A71 was still 17.0-fold (95% CI: 12.7- to 22.7-fold) higher than that of CVA16 patients (Supplemental Fig. 2).

The median and GMT of anti-CVA16 NAb responses were similar between mild and severe patients, with an overall difference close to 1.0-fold (95% CI: 0.6- to 1.7-fold), which persisted from 10 d to 26 mo (Fig. 4A). In general, the median and GMT of anti–EV-A71 NAb responses in severe patients were significantly higher than those of mild patients, with a 2.0-fold (95% CI: 1.4- to 3.2-fold) increase (Fig. 4B). These trends were coincident with those using CNS complications or PICU as a proxy for severity or in subgroup analysis after excluding the IVIG treatment effect.

FIGURE 4.

Clinical severity-specific NAb dynamics after illness onset of CVA16 and EV-A71 infection. (A) CVA16 infection. (B) EV-A71 infection. Severe patients who were admitted to the PICU or had CNS complications are shown. Black dotted lines indicate the GMT of NAb. The number of serum samples is listed below each time point. The Wilcoxon rank sum test was used to test the difference between mild and severe status. *p < 0.05, by t test. Note that 336 CVA16-infected patients (mild, 320; severe, 16) and 243 EV-A71–infected patients (mild, 144; severe, 99) were used in this comparison.

FIGURE 4.

Clinical severity-specific NAb dynamics after illness onset of CVA16 and EV-A71 infection. (A) CVA16 infection. (B) EV-A71 infection. Severe patients who were admitted to the PICU or had CNS complications are shown. Black dotted lines indicate the GMT of NAb. The number of serum samples is listed below each time point. The Wilcoxon rank sum test was used to test the difference between mild and severe status. *p < 0.05, by t test. Note that 336 CVA16-infected patients (mild, 320; severe, 16) and 243 EV-A71–infected patients (mild, 144; severe, 99) were used in this comparison.

Close modal

The acute phase of CVA16 infection was defined as <5.4 d according to the average period of disappearance of typical acute symptoms. The convalescent periods with seroconversion, a 4-fold increase, or an 8-fold increase were prolonged, as the sera were collected during the acute phase. Compared to daily sampling in the acute phase, the corresponding convalescent phases with seroconversion as early as possible ranged from 3.4 d (95% CI: 3.0–3.9 d) to 6.9 d (95% CI: 5.8–9.7 d) after illness onset. Similarly, the earliest convalescent phases with at least a 4-fold increase and an 8-fold increase were from 3.4 d (95% CI: 2.9–3.8 d) to 17.4 d (95% CI: 9.6–27.4 d) and from 3.7 d (95% CI: 3.3–4.2 d) to 17.2 d (95% CI: 6.0–24.6 d) after illness onset, respectively (Table III). Sensitivity analyses of convalescent time estimations are shown in the Supplemental Tables I and II and indicated similar results to the former values.

Table III.

Comparison of serum sampling time in CVA16- and EV-A71–infected HFMD patients during the acute and convalescent phases

Acute Phase (day)Convalescent Phase (day) (mean, 95% CI)
Seroconversion4-Fold Increase8-Fold Increase
CVA16CVA16EV-A71aDifferenceCVA16EV-A71aDifference
3.4 (3.0, 3.9) 3.4 (2.9, 3.8) 2.9 (2.6, 3.3) 0.5 3.7 (3.3, 4.2) 3.6 (3.1, 4.0) 0.1 
4.7 (4.1, 5.5) 3.8 (3.5, 4.2) 4.2 (3.6, 5.1) 0.4 5.2 (4.0, 7.0) 5.9 (4.3, 9.5) −0.7 
5.5 (4.1, 6.7) 6.3 (4.6, 8.5) 10.1 (6.3, 17.2) −3.8 7.6 (5.5, 10.5)b 16.4 (9.9, 26.9)b −8.8 
6.4 (5.4, 8.5) 12.9 (7.8, 20.7)b 14.9 (9.1, 23.8)b −2.0 13.7 (6.0, 20.7)b 21.4 (12.0, 31.9)b −7.7 
6.9 (5.8, 9.7) 17.4 (9.6, 27.4)b 24.4 (15.3, 38.3)b −7.0 17.2 (6.0, 24.6)b 22.9 (13.0, 31.6)b −5.7 
Acute Phase (day)Convalescent Phase (day) (mean, 95% CI)
Seroconversion4-Fold Increase8-Fold Increase
CVA16CVA16EV-A71aDifferenceCVA16EV-A71aDifference
3.4 (3.0, 3.9) 3.4 (2.9, 3.8) 2.9 (2.6, 3.3) 0.5 3.7 (3.3, 4.2) 3.6 (3.1, 4.0) 0.1 
4.7 (4.1, 5.5) 3.8 (3.5, 4.2) 4.2 (3.6, 5.1) 0.4 5.2 (4.0, 7.0) 5.9 (4.3, 9.5) −0.7 
5.5 (4.1, 6.7) 6.3 (4.6, 8.5) 10.1 (6.3, 17.2) −3.8 7.6 (5.5, 10.5)b 16.4 (9.9, 26.9)b −8.8 
6.4 (5.4, 8.5) 12.9 (7.8, 20.7)b 14.9 (9.1, 23.8)b −2.0 13.7 (6.0, 20.7)b 21.4 (12.0, 31.9)b −7.7 
6.9 (5.8, 9.7) 17.4 (9.6, 27.4)b 24.4 (15.3, 38.3)b −7.0 17.2 (6.0, 24.6)b 22.9 (13.0, 31.6)b −5.7 

The acute phase and convalescent phase both refer to time since the illness onset date.

a

Data were similar to those described in our previous study (22). We excluded the EV-A71 and CVA16 both-positive patients for comparison in this study.

b

Latency to collection after the average discharge time.

Using the same standard of infection, 4- and 8-fold increases, the difference in the second serum collection time between CVA16 and EV-A71 patients (22) was <1 d when the first serum was collected within 2 d (Table III). When the first serum was collected on the third to fifth day after onset, the second serum collection time for CVA16 was 2.0–8.8 d earlier than that for EV-A71 (22).

Our study found that CVA16 infection induced a persistent humoral immune response over 2 y, which was similar to that of EV-A71 infection (22) in HFMD patients, but the CVA16-induced immune response was significantly lower than that induced by EV-A71. Compared with those with mild disease, there was a trend for severely ill patients to exhibit an increase in the anti–EV-A71 NAb response over time, but this trend was not significant in CVA16-infected patients. Thus, it is recommended that the acute phase sample for serological diagnosis of CVA16-infected patients should be taken as early as possible after illness onset, ideally <3 d, and then convalescent sera can be collected before discharge.

Our data showed that the seropositivity (78.6%) and GMT (111; 95% CI: 37–330) of the NAb response of the EV-A71–infected patients was higher at 1 d after illness onset, which are similar values to those reported in previous studies (11, 22, 27). In addition, the Lin et al. (21) study also showed a high positive percentage of anti–EV-A71 IgM (87.5%) within 1–3 d of illness onset among EV-A71 RNA-positive HFMD patients. In contrast to that of EV-A71–infected patients, the seropositive percentage of CVA16-infected patients was still <50% within 3 d of illness onset, but anti-NAb of the two viruses both reached 100% seropositivity in the follow-up. A longitudinal study of EV-A71– or CVA16-infected patients conducted in Vietnam showed differences in the seropositive percentage and GMT between anti–EV-A71 NAb and anti-CVA16 NAb. Their results showed that the seropositive percentage changed from 60% for EV-A71 and 27% for CVA16 at enrollment (<3 d after onset) to 100% at the second serum collection (within 7–14 d after enrollment), and the corresponding GMTs of EV-A71 patients increased from 38 to 295 and were lower for CVA16 (which increased from 10 to 141), similar to our results (11). The differences in the positive percentage of IgM during the course of infection with EV-A71 and CVA16 were also consistent with our finding of an increase in the seropositive percentage (17). In addition, the kinetics for the anti-CVA16 IgM response decreased after 20 wk, similar to those in our study.

Although our results showed that CVA16 infections can effectively induce a long-term immune response similar to that induced by EV-A71 (22), the GMT increase was relatively slow, and the NAb response level was significantly lower than that for EV-A71 (peak time, 21 versus 13 d; peak titer, 305 versus 2417). After adjusting for age and clinical severity, the number of EV-A71–infected patients was still higher than that of CVA16 patients. These results suggest that EV-A71 might elicit stronger NAb responses than CVA16, and a similar pattern was also observed in previous studies (11, 12, 17). This difference is probably due to the ∼20% sequence difference in the capsid proteins of the two viruses and the antigenic variation caused by the different surface structures of the two viruses (28).

In another study (J. Chen, K. Wang, Y. Zhou, Y. Cheng, L. Liang, Q. Qiu, P. Cui, Y. Li, L. Turtle, and H. Yu, unpublished observations), we found that NAb titers in EV-A71–infected patients with CNS complications tended to be higher than those in patients without CNS complications, but these data were not analyzed for longer than a 2-mo duration. There were no previous longitudinal data on the differences in the kinetics of the anti-CVA16 NAb response level between mild and severe HFMD patients. For severe HFMD patients, IVIG appeared to be beneficial as it shortened the duration of illness episodes in a previous report (5). IVIG is also recommended by several national and international guideline committees (29, 30). Our previous study showed that IVIG had no significant effect on the NAb response level or rate of GMT increase (22), but we need to prevent clinical severity from influencing the analysis. Thus, we further analyzed whether the clinical severity affected the NAb response level and whether IVIG affected the results in this study. Our results showed that the anti–EV-A71 NAb response level of severely ill patients who developed CNS complications or required PICU admission showed a tendency to increase during 2 y; this response was significantly higher than that of patients with mild disease only at 6 d after illness onset. However, the power of this analysis was not sufficient to reveal whether the anti-CVA16 NAb response level differed in patients with severe disease and those with mild disease due to the limited severe case sample size, especially in the long-term follow-up.

Despite no worldwide consensus for paired-serum sampling time in HFMD serological diagnosis, convalescent sera should be collected before case discharge from an optimal practical standpoint. Thus, the standard for collecting the acute sera of CVA16-infected patients should be within 3 d according to our results, which is also consistent with Chinese HFMD laboratory manual recommendations (31) and most previous studies (11, 14, 15). Moreover, our results suggest that the corresponding convalescent serum of patients infected with CVA16 needs to be collected no less than 17.4 d (95% CI: 9.6–27.4 d) after onset based on at least a 4-fold increase, which is also similar to the Shi et al. (16) study (i.e., 14–30 d). In the Lin et al. (21) study, the titers of almost 94.6% of paired sera had increased at least 4-fold when collecting paired samples with an interval of 4–7 d, which is in line with our recommendation. However, the collection time of convalescence samples from other studies (11, 14, 19, 20) lagged in comparison with the shortest interval (2.4 d) and the longest intervals (12.4 d) in our studies. Thus, prior studies may not have collected the timeliest samples, and too long of a time interval between the collection of paired samples makes diagnosis slow and often clinically irrelevant.

In our study, we used the same standard of infection such as a 4- or 8-fold increase for the first serum sample (collected on the third to fifth day after onset) and the corresponding second serum sample, but the collection time for these samples from CVA16-infected patients was 2.0–8.8 d earlier than that for EV-A71 (22). This may be due to the negative GMT and slower increase in the initial stage of CVA16-infected patients, but the anti–EV-A71 NAb titer rapidly increased to relatively high GMT around the peak titer, posing difficulty in determining the corresponding convalescent sera with at least a 4- or 8-fold increase.

Our study had some limitations. First, the low follow-up rate (CVA16 22%, EV-A71 42%) and sparse data on the timing of serum collection might have caused deviation in the model, especially in the long-term estimation. Second, our study was conducted in only a single center, and the results may not be suitable for extrapolation to other regions. The research results need to be further verified. Third, we cannot conclude whether the clinical severity had an effect on the anti-CVA16 NAb response level, as too few severe CVA16-infected patients were enrolled in our study. The date with the maximum GMT was slightly different between the model prediction and observed GMT in our study, which was probably due to the model excluding some patients and considering the autocorrelation for patients. The slight waning in NAb response after 4.5 mo may be limited by the relatively small sample size, although there was a similar pattern in anti-CVA16 IgM studies. The immune response to CVA16 was mild without an extreme increase; thus, the recovery phases that used 4- and 8-fold increases were similar.

In conclusion, to our knowledge, our study is the first to describe the kinetics of the anti-CVA16 NAb response during hospitalization and for up to 26 mo after recovery and to compare the kinetics of the anti–EV-A71 NAb response by using data from the same prospective cohort of infected HFMD patients. We found that CVA16 infection induced relatively low levels of NAb titers; they were maintained above the protective NAb titer threshold until at least 2 y after illness onset, similar to EV-A71 infection. In addition, the clinical severity had an effect on the anti-EV-A71 NAb response level. In addition, the ideal time of sampling paired sera for serological diagnosis of CVA16- and EV-A71–infected patients differed. The acute phase sample should be taken as early as possible (ideally, <3 d) after illness onset, and convalescent sera can be collected before discharge for both infections.

We thank the staff members at Henan Children’s Hospital for assisting with the field investigation, administration, and data collection.

This work was supported by National Science Fund for Distinguished Young Scholars Grant 81525023.

The online version of this article contains supplemental material.

H.Y. designed and supervised the study. Jianli Yang, Junmei Yang, Y.L., L.L., P.C., Y.C., and C.Z. conducted the investigation and collected the samples and data. Y.Z., Q.Q., L.W., W.Z., H.S., H.G., and K.W. conducted the experiments. J.Z. and Y.Z. analyzed the data. Y.Z. and J.Z. wrote the initial drafts of the manuscript. H.Y. and J.J.H.C. commented on and revised drafts of the manuscript. All authors contributed to the review and revision, approved the final manuscript as submitted, and agree to be accountable for all aspects of the work.

Abbreviations used in this article:

     
  • CI

    confidence interval

  •  
  • CVA16

    coxsackievirus A16

  •  
  • EV-A71

    enterovirus A71

  •  
  • GLM

    generalized linear model

  •  
  • GMT

    geometric mean titer

  •  
  • HFMD

    hand, foot, and mouth disease

  •  
  • IQR

    interquartile range

  •  
  • IVIG

    i.v. Ig

  •  
  • LOS

    length of hospital stay

  •  
  • NAb

    neutralizing Ab

  •  
  • PICU

    pediatric intensive care unit

1.
Yu
H.
,
B. J.
Cowling
.
2019
.
Remaining challenges for prevention and control of hand, foot, and mouth disease.
Lancet Child Adolesc. Health
3
:
373
374
.
2.
Xing
W.
,
Q.
Liao
,
C.
Viboud
,
J.
Zhang
,
J.
Sun
,
J. T.
Wu
,
Z.
Chang
,
F.
Liu
,
V. J.
Fang
,
Y.
Zheng
, et al
2014
.
Hand, foot, and mouth disease in China, 2008–12: an epidemiological study.
Lancet Infect. Dis.
14
:
308
318
.
3.
Fischer
C.
,
F.
Bozza
,
X. J.
Merino Merino
,
C.
Pedroso
,
E. F.
de Oliveira Filho
,
A.
Moreira-Soto
,
A.
Schwalb
,
X.
de Lamballerie
,
E. M.
Netto
,
P. T.
Bozza
, et al
2020
.
Robustness of serologic investigations for Chikungunya and Mayaro viruses following coemergence.
MSphere
5
:
e00915-19
.
4.
Takahashi
S.
,
C. J. E.
Metcalf
,
Y.
Arima
,
T.
Fujimoto
,
H.
Shimizu
,
H.
Rogier van Doorn
,
T.
Le Van
,
Y. F.
Chan
,
J. J.
Farrar
,
K.
Oishi
,
B. T.
Grenfell
.
2018
.
Epidemic dynamics, interactions and predictability of enteroviruses associated with hand, foot and mouth disease in Japan.
J. R. Soc. Interface
15
:
20180507
.
5.
Cai
K.
,
Y.
Wang
,
Z.
Guo
,
H.
Yu
,
H.
Li
,
L.
Zhang
,
S.
Xu
,
Q.
Zhang
.
2019
.
Clinical characteristics and managements of severe hand, foot and mouth disease caused by enterovirus A71 and coxsackievirus A16 in Shanghai, China.
BMC Infect. Dis.
19
:
285
.
6.
Li
Y.
,
Y.
Zhou
,
Y.
Cheng
,
P.
Wu
,
C.
Zhou
,
P.
Cui
,
C.
Song
,
L.
Liang
,
F.
Wang
,
Q.
Qiu
, et al
2019
.
Effectiveness of EV-A71 vaccination in prevention of paediatric hand, foot, and mouth disease associated with EV-A71 virus infection requiring hospitalisation in Henan, China, 2017–18: a test-negative case-control study.
Lancet Child Adolesc. Health
3
:
697
704
.
7.
Han
Z.
,
Y.
Song
,
J.
Xiao
,
L.
Jiang
,
W.
Huang
,
H.
Wei
,
J.
Li
,
H.
Zeng
,
Q.
Yu
,
J.
Li
, et al
2020
.
Genomic epidemiology of coxsackievirus A16 in mainland of China, 2000–18.
Virus Evol.
6
:
veaa084
.
8.
Du
Z.
,
Y.
Huang
,
W. R.
Lawrence
,
J.
Xu
,
Z.
Yang
,
J.
Lu
,
Z.
Zhang
,
Y.
Hao
.
2021
.
Leading enterovirus genotypes causing hand, foot, and mouth disease in Guangzhou, China: relationship with climate and vaccination against EV71.
Int. J. Environ. Res. Public Health
18
:
292
.
9.
Li
W.
,
L.
Yi
,
J.
Su
,
J.
Lu
,
C.
Ke
,
H.
Zeng
,
D.
Guan
,
C.
Ma
,
W.
Zhang
,
H.
Xiao
, et al
2013
.
Seroprevalence of human enterovirus 71 and coxsackievirus A16 in Guangdong, China, in pre- and post-2010 HFMD epidemic period.
PLoS One
8
:
e80515
.
10.
Mao
Q. Y.
,
X. Y.
Liao
,
X.
Yu
,
N.
Li
,
F. C.
Zhu
,
Y.
Zeng
,
Z. L.
Liang
,
F. X.
Li
,
J. Z.
Wang
,
F. M.
Lu
,
H.
Zhuang
.
2010
.
Dynamic change of mother-source neutralizing antibodies against enterovirus 71 and coxsackievirus A16 in infants.
Chin. Med. J. (Engl.)
123
:
1679
1684
.
11.
Nguyet
L. A.
,
T. T.
Thanh
,
L. N. T.
Nhan
,
N. T. T.
Hong
,
L. N. T.
Nhu
,
H. M. T.
Van
,
N. T. H.
Ny
,
N. T.
Anh
,
D. D. K.
Han
,
H. M.
Tuan
, et al
2020
.
Neutralizing antibodies against enteroviruses in patients with hand, foot and mouth disease.
Emerg. Infect. Dis.
26
:
298
306
.
12.
Gao
F.
,
Q. Y.
Mao
,
P.
Chen
,
L. L.
Bian
,
X.
Yao
,
J. X.
Li
,
F. C.
Zhu
,
Z. L.
Liang
.
2016
.
Seroepidemiology of coxsackievirus A6, coxsackievirus A16, and enterovirus 71 infections in infants and children: a prospective cohort study in Jiangsu, China.
J. Infect.
73
:
509
512
.
13.
Harvala
H.
,
E.
Broberg
,
K.
Benschop
,
N.
Berginc
,
S.
Ladhani
,
P.
Susi
,
C.
Christiansen
,
J.
McKenna
,
D.
Allen
,
P.
Makiello
, et al
2018
.
Recommendations for enterovirus diagnostics and characterisation within and beyond Europe.
J. Clin. Virol.
101
:
11
17
.
14.
Gao
F.
,
Y. P.
Wang
,
Q. Y.
Mao
,
X.
Yao
,
S.
Liu
,
F. X.
Li
,
F. C.
Zhu
,
J. Y.
Yang
,
Z. L.
Liang
,
F. M.
Lu
,
J. Z.
Wang
.
2012
.
Enterovirus 71 viral capsid protein linear epitopes: identification and characterization.
Virol. J.
9
:
26
.
15.
Yan
X.
,
Z. Z.
Zhang
,
Z. H.
Yang
,
C. M.
Zhu
,
Y. G.
Hu
,
Q. B.
Liu
.
2015
.
Clinical and etiological characteristics of atypical hand-foot-and-mouth disease in children from Chongqing, China: a retrospective study.
BioMed Res. Int.
2015
:
802046
.
16.
Shi
M.
,
Y.
Zhou
,
L.
Cao
,
C.
Ding
,
Y.
Ji
,
Q.
Jiang
,
X.
Liu
,
X.
Li
,
X.
Hou
,
H.
Peng
,
W.
Shi
.
2013
.
Expression of enterovirus 71 capsid protein VP1 in Escherichia coli and its clinical application.
Braz. J. Microbiol.
44
:
1215
1222
.
17.
Yu
N.
,
M.
Guo
,
S. J.
He
,
Y. X.
Pan
,
X. X.
Chen
,
X. X.
Ding
,
W.
Hao
,
Y. D.
Wang
,
S. X.
Ge
,
N. S.
Xia
,
X. Y.
Che
.
2012
.
Evaluation of human enterovirus 71 and coxsackievirus A16 specific immunoglobulin M antibodies for diagnosis of hand-foot-and-mouth disease.
Virol. J.
9
:
12
.
18.
Zhang
A.
,
B.
Xiu
,
H.
Zhang
,
N.
Li
.
2016
.
Protein microarray-mediated detection of antienterovirus antibodies in serum.
J. Int. Med. Res.
44
:
287
296
.
19.
Dang
S.
,
N.
Gao
,
Y.
Li
,
M.
Li
,
X.
Wang
,
X.
Jia
,
S.
Zhai
,
X.
Zhang
,
J.
Liu
,
H.
Deng
,
T.
Dong
.
2014
.
Dominant CD4-dependent RNA-dependent RNA polymerase-specific T-cell responses in children acutely infected with human enterovirus 71 and healthy adult controls.
Immunology
142
:
89
100
.
20.
Li
C. C.
,
M. Y.
Yang
,
R. F.
Chen
,
T. Y.
Lin
,
K. C.
Tsao
,
H. C.
Ning
,
H. C.
Liu
,
S. F.
Lin
,
W. T.
Yeh
,
Y. T.
Chu
,
K. D.
Yang
.
2002
.
Clinical manifestations and laboratory assessment in an enterovirus 71 outbreak in southern Taiwan.
Scand. J. Infect. Dis.
34
:
104
109
.
21.
Lin
Y.
,
K.
Wen
,
Y.
Pan
,
Y.
Wang
,
X.
Che
,
B.
Wang
.
2011
.
Cross-reactivity of anti-EV71 IgM and neutralizing antibody in series sera of patients infected with enterovirus 71 and coxsackievirus A 16.
J. Immunoassay Immunochem.
32
:
233
243
.
22.
Qiu
Q.
,
J.
Zhou
,
Y.
Cheng
,
Y.
Zhou
,
L.
Liang
,
P.
Cui
,
Y.
Xue
,
L.
Wang
,
K.
Wang
,
H.
Wang
, et al
2021
.
Kinetics of the neutralising antibody response in patients with hand, foot, and mouth disease caused by EV-A71: a longitudinal cohort study in Zhengzhou during 2017–2019.
EBioMedicine
68
:
103398
.
23.
World Health Organization
.
2011
.
A Guide to Clinical Management and Public Health Response for Hand, Foot and Mouth Disease (HFMD).
World Health Organization
,
Geneva
.
24.
Zhu
R.
,
T.
Cheng
,
Z.
Yin
,
D.
Liu
,
L.
Xu
,
Y.
Li
,
W.
Wang
,
J.
Liu
,
Y.
Que
,
X.
Ye
, et al
2018
.
Serological survey of neutralizing antibodies to eight major enteroviruses among healthy population.
Emerg. Microbes Infect.
7
:
2
.
25.
World Health Organization
.
2004
.
Polio Laboratory Manual
, 4th Ed.
Available at: https://apps.who.int/iris/handle/10665/68762. Accessed June 23, 2017
.
26.
Wei
X.
,
J.
Yang
,
L.
Gao
,
L.
Wang
,
Q.
Liao
,
Q.
Qiu
,
K.
Luo
,
S.
Yu
,
Y.
Zhou
,
F.
Liu
, et al
2021
.
The transfer and decay of maternal antibodies against enterovirus A71, and dynamics of antibodies due to later natural infections in Chinese infants: a longitudinal, paired mother-neonate cohort study.
Lancet Infect. Dis.
21
:
418
426
.
27.
Yang
C.
,
C.
Deng
,
J.
Wan
,
L.
Zhu
,
Q.
Leng
.
2011
.
Neutralizing antibody response in the patients with hand, foot and mouth disease to enterovirus 71 and its clinical implications.
Virol. J.
8
:
306
.
28.
Ren
J.
,
X.
Wang
,
L.
Zhu
,
Z.
Hu
,
Q.
Gao
,
P.
Yang
,
X.
Li
,
J.
Wang
,
X.
Shen
,
E. E.
Fry
, et al
2015
.
Structures of coxsackievirus A16 capsids with native antigenicity: implications for particle expansion, receptor binding, and immunogenicity.
J. Virol.
89
:
10500
10511
.
29.
Wang
J.
,
Z.
Teng
,
X.
Cui
,
C.
Li
,
H.
Pan
,
Y.
Zheng
,
S.
Mao
,
Y.
Yang
,
L.
Wu
,
X.
Guo
, et al
2018
.
Epidemiological and serological surveillance of hand-foot-and-mouth disease in Shanghai, China, 2012–2016.
Emerg. Microbes Infect.
7
:
8
.
30.
Yen
M. H.
,
Y. C.
Huang
,
M. C.
Chen
,
C. C.
Liu
,
N. C.
Chiu
,
R.
Lien
,
L. Y.
Chang
,
C. H.
Chiu
,
K. C.
Tsao
,
T. Y.
Lin
.
2015
.
Effect of intravenous immunoglobulin for neonates with severe enteroviral infections with emphasis on the timing of administration.
J. Clin. Virol.
64
:
92
96
.
31.
Li
X. W.
,
X.
Ni
,
S. Y.
Qian
,
Q.
Wang
,
R. M.
Jiang
,
W. B.
Xu
,
Y. C.
Zhang
,
G. J.
Yu
,
Q.
Chen
,
Y. X.
Shang
, et al
2018
.
Chinese guidelines for the diagnosis and treatment of hand, foot and mouth disease (2018 edition).
World J. Pediatr.
14
:
437
447
.

H.Y. has received research funding from Sanofi Pasteur, GlaxoSmithKline, Yichang HEC Changjiang Pharmaceutical Company, and Shanghai Roche Pharmaceutical Company, outside of this study. The other authors have no financial conflicts of interest.

Supplementary data