Hand, foot, and mouth disease (HFMD) is a common viral disease that afflicts children. The symptoms of the disease are generally mild, self-limiting, and characterized by fever, herpangina and rash on extremities. A few patients may develop severe complications involving the nervous system, such as aseptic meningitis, brain stem encephalitis, acute flaccid paralysis, and neurogenic pulmonary edema, or even death [1]. Transmission of the virus via direct person-to-person contact or indirect transmission through contaminated objects leads to frequent outbreaks in crowded settings. In the late 1990s, Taiwan, Malaysia, and Singapore experienced large-scale outbreaks of HFMD [2–4]. In mainland China, HFMD was sporadically reported only during the 1980s and 90s [5–7]. No large-scale outbreaks were observed until sudden local outbreaks occurred in Shandong and Anhui provinces in 2007-2008 [8, 9]. The outbreaks and the subsequent widespread occurrence of HFMD posed a serious threat to children’s health. HFMD was categorized as a notifiable infectious disease on May 2, 2008, in mainland China. Comprehensive and enhanced surveillance of the disease has been implemented nationwide since then. Systematic surveillance in recent years has revealed that HFMD is widespread in mainland China, and millions of cases, including hundreds of deaths, are reported annually.

HFMD is caused by human enteroviruses (EVs) belonging to the genus Enterovirus of the family Picornaviridae. The EV virion is composed of a capsid and a positive single-strand RNA genome whose sequence undergoes frequent variation. According to their antigenic properties and pathogenicity [10], EVs have been divided into numerous serotypes and grouped traditionally into poliovirus (PV, 3 serotypes), coxsackie virus A (CVA, 23 serotypes), coxsackievirus B (CVB, 6 serotypes), and echovirus (ECHO, 28 serotypes). Advances in molecular typing have led to the classification of enteroviruses into four species: A to D [11]. Genetically, the viruses with identical serotypes have been further divided into multiple genotypes or sub-genotypes. Among the multiple serotypes of EV, EV71 and CVA16 are the two major serotypes associated with HFMD. In addition to the two main serotypes, infection with multiple EVs (non-EV71 and non-CVA16) and such as CVA2, 4, 5, 6, 8, 10 and 16, CVB1-5, also results in HFMD pathogenesis with diverse clinical manifestations. The diversity of serotypes associated with HFMD and the occurrence of genetic mutations are regarded as important factors contributing to the HFMD epidemic. Co-circulation of multiple serotypes and genetic variation of EV contribute to the complex epidemiology of HFMD.

In addition to etiological factors, natural factors are thought to be involved in the HFMD epidemic. Previous studies have shown that climate and geographical location are important factors in the HFMD epidemic in mainland China [12]. The abundance of rainfall and higher average temperature in southern areas compared with the northern areas contributes to the higher incidence of HFMD [13–15]. Fujian is a coastal province located in southeastern China, with a typical sub-tropical monsoon climate. In recent years, the incidence of HFMD in the province has increased annually. However, it remains unclear whether the HFMD epidemic in Fujian is similar to that in southern China. Understanding the HFMD epidemic is important for effective prevention and control of the disease. We therefore investigated the HFMD epidemic in Fujian using descriptive epidemiology combined with laboratory monitoring.

This study was approved by the Ethics Review Committee of the Fujian Provincial Center for Disease Control and Prevention (CDC). Specimens were collected from patients who underwent regular medical examination at sentinel hospitals, and therefore, written informed consent was waived. All of the information collected from patients, including demographic data, clinical records, and laboratory findings was kept anonymous to protect their privacy.

The records for reported HFMD cases were retrieved according to the date of onset of illness from May 2, 2008, to December 31, 2014, and downloaded from the National Disease Surveillance Information Report Administration System of China. The definitions of mild, severe and fatal cases were adopted according to the Guidelines for response to hand-foot-and-mouth disease epidemic (http://www.chinacdc.cn/jkzt/crb/bl/szkb/jszl_2275/200906/t20090612_24707.html ) issued by the China CDC. Epidemiological characteristics including temporal and spatial distribution, demographic features, and severity of illness were descriptively analyzed. Data sorting and statistics were performed using Microsoft Excel. The correlation between the serotype and severity of illness was analyzed by linear regression using GraphPad Prism (Ver. 5.0), and the geographic distribution of the disease was analyzed using ArcGIS (Ver.10.2).

To obtain the viral isolates for etiological characterization of HFMD, throat and/or cloacal swab specimens were collected from HFMD patients through the surveillance network of sentinel hospitals and laboratories at the municipal CDCs of Fujian. The swabs were transferred to 3-5 mL of DMEM (Gibco) containing 5% FBS (Pan Biotech) and delivered to laboratories within 12 h for molecular diagnosis using one-step RT-PCR or real-time one-step RT-PCR according to the Guidelines for response to hand-foot-and-mouth disease epidemic. The positive specimens were inoculated in parallel on human rhabdomyosarcoma cells (RD, ATCCRCCL-136) and human epidermoid cancer cells (HEp-2, ATCCRCCL-23) for viral isolation. Cellular infection and viral propagation were confirmed by observing cytopathic effects. One-step RT-PCR was used to amplify the RNA extracted from the supernatant with the following universal primers for EV detection: PE2 (5’-TCCGGCCCCTGAATGCGGCTAATCC-3’) and PE1 (5’-ACACGGACACCCAAAGTAGTCGGTCC-3’).

To analyze the genetic variation of the main serotypes of EVs, 179 strains of EV71 and 60 strains of CVA16 isolated from HFMD patients with various clinical outcomes were randomly selected during 2010-2014. Complete VP1 gene amplification and sequencing were performed using the EV71-specific primers EV71/VP1F (5’-CAGGAAACAGCTATGACGCAGCCCAAAAGAACTTCAC-3’)

and EV71/VP1R (5’-AGTCACGACGTTGTAAAGTCGCGAGAGCTGTCTTC-3’) and the CVA16-specific primers CVA16/VP1F(5’-CAGGAAACAGCTATGACAGCCCAGGACAACTTCAC-3’) and CVA16/VP1R (5’-AGTCACGACGTTGTAGCCGATTCACTACCCTAT-3’). After agarose gel separation and column purification, all of the amplicons were subjected directly to bidirectional sequencing on a Genetic Analyzer 3500 platform (Life Technologies) with M13-RV and M13-M primers fused to the 5’ termini of forward and reverse primers (underlined), respectively. All of the viral sequences obtained in this study were deposited in the GenBank database (accession nos. KU595714-595892 and KU595958-596017). Together with the sequences of genotype/subgenotype reference strains and viral isolates found in other provinces of China, VP1 sequences obtained in this study were aligned using MEGA6.0 software (Ver. 6.0). Phylogenetic trees of EV71 and CVA16 were reconstructed by the neighbor-joining method (Kimura 2-parameter model) and validated by bootstrap tests (1,000 repeats).

To determine the serotype spectrum of other EVs (non-EV71 and non-CVA16) in Fujian, partial VP1 fragments obtained from 407 strains of other EVs from HFMD patients during 2011-2014 were amplified by one-step RT-PCR with various primers as reported previously [16, 17]. The amplicons were gel-purified and subsequently inserted into pMD20T vector (TaKaRa) by A-T cloning. The recombinant plasmids were purified and subjected to bidirectional sequencing using M13-RV and M13-M4 primers flanking the insertion site. All of the sequences of other EVs were submitted to the online Enterovirus Genotyping Tool (http://www.rivm.nl/mpf/enterovirus/typingtool) for viral serotyping.

A total of 448,788 cases, including 2,211 severe cases and 88 deaths, were reported in Fujian between May 2, 2008 and Dec 31, 2014. Overall, the incidence rates of HFMD increased from 4.77 to 30.14 per 10,000 annually between 2008 and 2014. However, the case-severity, case-fatality and severity-fatality rates decreased after 2010 (Table S1, Supplementary Material). The temporal distribution indicated that the HFMD epidemic in Fujian had clear seasonality. The incidence peaked in May, with a secondary peak in September each year (Fig. 1). However, the distinct temporal distribution between EV71 and CVA16 reported previously [18] was not observed in Fujian. The spatial distribution indicated that HFMD was spread across nine cities and one pilot zone in the province, with incidence rates varying between subregions during the period of 2010-2014. Overall, the incidence rates in coastal areas were slightly higher than in inland areas (Figure S1, Supplementary Material). We speculate that this might be related to the higher population density and abundant rainfall in coastal areas [13]. The demographic distribution (Table S2) indicated that males outnumbered females, with a ratio of 1.72 (1.60-1.87, male/female). Among the reported cases, children younger than 4 years accounted for 56.15% (322,478 cases) and the 1-year age group was the most affected (149,220 cases, 33.25%). Individual children accounted for 77.96%, followed by kindergarten children, who accounted for 19.22%. Etiological monitoring revealed that the serotype distribution of EVs co-circulating in Fujian changed greatly during 2008-2014. In brief, EV71 was predominant during 2008-2009 and tended to decrease after 2010 despite its predominance until the third quarter of 2012. EV71 was replaced by other EVs after the fourth quarter of 2012, except for the period from the fourth quarter of 2013 to the first half of 2014, during which CVA16 was most prevalent (Fig. S2).

Fig. 1
figure 1

Weekly distribution of HFMD in Fujian, 2008-2014. Case registries were collected and sorted by the date of disease onset since May 2, 2008

Full size image

The case-severity and case-fatality rates of HFMD in Fujian varied each year. The rates peaked in 2010 and declined subsequently (Table S1). The correlation between disease severity and EV serotype revealed that EV71 was the leading etiological agent associated with severity and fatality of HFMD cases in Fujian. During 2010-2014, EV71 infection accounted for 40.23% (5,648/14,040) of mild cases. However, the proportion of EV71 infections increased significantly to 72.8% (1,194/1,640) and 98.59% (70/71) in severe and fatal cases, respectively. Linear regression analysis revealed that the incidence of EV71 infections was 75.74% (2,548/3,364), 43.88% (624/1,422), 50.19% (1,979/3,943), 23.65% (808/3,417) and 26.44% (953/3,605) each year from 2010 to 2014, which was strongly correlated with annual case-severity rates of HFMD (r 2 = 0.9254, p = 0.0089). By contrast, CVA16 infection mainly caused mild HFMD (22.93%, 3,220/14,040) and accounted only for 7.01% (115/1,640) of severe cases, with no deaths. Other EVs also caused mild HFMD (36.84%, 5,172/14,040) and accounted for 20.18% of severe cases (331/1,640) and 1.41% of all of the deaths (1/71). Notably, the case-severity and case-fatality rates in 2008-2009, compared with those in 2010-2014, were relatively low in spite of the higher incidence of EV71 (Table S1). The underlying causes included neglect of most patients due to the relatively mild manifestation of other EVs and poor surveillance networks in the initial stages of enhanced disease surveillance. The quality of reporting was improved by adjusting the surveillance strategy of China CDC as well as development of diagnostic methods for other EVs.

VP1-based phylogenetic analysis of EV71 and CVA16 was conducted to investigate the phylogenesis of the two main serotypes of EV71 and CVA16 in this study. The reconstructed phylogenetic tree indicated that all of the EV71 viruses circulating in Fujian during 2010-2014 belonged to clade C4a and were not clustered by clinical severity or time of appearance (Fig. 2A). The phylogenetic tree of CVA16 indicated that the viruses circulating in Fujian during the same period belonged to two clades, B1a and B1b. Similar to EV71, the viruses of CVA16 showed no clustering by clinical severity or time of appearance (Fig. 2B). Genetically, EV71 was classified into genotypes A, B and C, and the latter two were further divided into subgenotypes B1-5 and C1-5. Molecular epidemiology suggested that the two clades, C4a and C4b, of subgenotype C4 of EV71 were associated with a persistent epidemic in mainland China [19], and the clade shift from C4b to C4a occurred around 2004 [20]. CVA16 comprised genotypes A and B, and the latter included two subgenotypes, B1 and B2. Two clades in the B1 subgenotype (B1b and B1a) were associated with the HFMD epidemic and circulated in mainland China [21]. Therefore, the EV71 and CVA16 circulating in Fujian were genetically similar to those circulating synchronously nationwide. Despite the circulation of many genetically heterogenous viral strains, no viral strain of EV71 or CVA16 demonstrated superior dissemination because no temporal changes in severity or incidence were observed.

Fig. 2
figure 2figure 2

VP1-based phylogenetic analysis of EV71 (A) and CVA16 (B) in Fujian, 2010-2014. The viruses isolated in 2010, 2011, 2012, 2013, and 2014 are colored brown, green, purple, blue and red, respectively. Viruses isolated from mild (M), severe (S), and fatal (F) cases, are shown. Bootstrap values greater than 95% are displayed above the nodes. Reference sequences of various genotypes or subgenotypes downloaded from GenBank and the corresponding accession numbers are indicated with the virus designation

Full size image

The spread of other EV (non-EV71 and non-CVA16) infections among HFMD patients has been reported in other regions of mainland China [22, 23]. Etiological monitoring revealed a similar phenomenon in Fujian in recent years (Fig. S2). We identified 22 serotypes of EV in Fujian based on partial VP1 sequences derived from 407 strains of other EVs isolated from HFMD patients during 2011-2014, including nine serotypes of CVA, five of CVB, seven of ECHO and one of EV68 (Table 1). Among the other EVs, CVA6 and CVA10 accounted for 48.89% (199/407) and 25.31% (103/407), respectively. The serotypes of other predominant EVs were CVA10 (41.56%, 32/77) in 2011 and CVA6 since 2012 (48.89-60.22%).

Table 1 Summary of other enterovirus serotypes detected in HFMD patients reported in Fujian, 2011-2014
Full size table

Based on genetic differences, more than 100 serotypes of EV have been identified so far [24]. HFMD is caused by multiple EV serotypes that cause diverse clinical manifestations [25, 26], which can differ even between similar EV serotypes. For example, an outbreak of viral encephalitis caused by ECHO30, which is difficult to distinguish from HFMD, was reported previously in Fujian. We postulated that co-circulation of multiple serotypes of EVs and the lack of cross-immunity has resulted in a considerable increase of HFMD in Fujian since 2011. Co-circulation of multiple EV serotypes is a diagnostic challenge, especially for atypical HFMD cases [26], as well as disease management and control.

Rapid advances in vaccine development against EV71 have been reported in mainland China, Taiwan and Singapore [27]. The inactivated EV71 vaccine derived from clade C4a of the C4 subgenotype, which was predominant in mainland China, showed efficient protection against viral infection and partial cross-protection against viruses of other genotypes [28–30]. Fortunately, two EV71 vaccines developed by different institutions were licensed toward the end of 2015 in mainland China. Considering that EV71 is the leading cause of severe HFMD, affordable vaccination has the potential to greatly reduce the severity and fatality of HFMD.

In summary, we investigated the epidemiology of HFMD in Fujian during the period of 2008-2014. The temporal and spatial distribution, demographics, and correlation between severity of illness and viral serotype in Fujian were consistent with those of an HFMD epidemic in the southern areas of China [31]. Nonetheless, the increasing incidence of HFMD, co-circulation of multiple serotypes, potential genetic variation among EVs similar to seasonal influenza viruses [32], and underlying epidemiological factors are still challenges for disease prevention and control. Continuous epidemiological and etiological surveillance of HFMD is, therefore, imperative.