OBJECTIVE. In 1998, an epidemic of hand-foot-mouth disease/herpangina was caused by human enterovirus 71 infection in Taiwan. The underlying factors of widespread emergence of viral infection are unclear. The purpose of this study was to assess the epidemiology of hand-foot-mouth disease/herpangina in Taiwan between March 1998 and December 2005.
METHODS. We analyzed data reported to surveillance systems at the Taiwan Center for Disease Control. Viral isolation was performed by 11 reference virus laboratories at medical centers as well as the Taiwan Center for Disease Control.
RESULTS. During the 8-year study period, the reported incidence of mild cases of hand-foot-mouth disease/herpangina varied from 0.8 to 19.9 cases per sentinel physician per week, peaking in 1998. Seasonal variations in incidence were observed, with an incidence peak observed during the summer season. Annual incidence changed significantly from 1998 to 2005. Both age-specific incidence and fatality of severe hand-foot-mouth disease/herpangina decreased as age increased. Most (93%) cases occurred in children who were aged 4 years and younger. Inpatients had a higher rate of enterovirus 71 infection than outpatients. Among severe cases, the majority (80%) had pulmonary edema/hemorrhage and encephalitis.
CONCLUSIONS. Hand-foot-mouth disease/herpangina is a common disease in Taiwan. Enterovirus 71 infection has emerged as an important public problem causing serious clinical illness and, potentially, death in young children. Vaccine development is recommended for prevention of enterovirus 71 infection in the future.
Hand-food-mouth disease (HFMD) is a common enteroviral infection, most frequently caused by coxsackievirus A16 and enterovirus 71 (EV 71).1,2 Young children are the most frequent shedders of enterovirus and are usually the index case during family outbreaks.3 EV 71 infection can result in aseptic meningitis, encephalitis, myocarditis, or poliomyelitis-like paralysis.4 Since the first record of EV 71 infection in California in 1969,5 subsequent outbreaks have been reported worldwide. In the early outbreaks during 1969–1974, serious central nervous system complication was uncommon.6,7 However, subsequent reports have revealed that the disease spectrum ranged from febrile disease, HFMD, herpangina, aseptic meningitis, polio-like paralysis, encephalitis, and even death.8 Although mild diseases are the predominant clinical features of EV 71 infection, neurologic involvement is the most serious complication and accounts for the remarkable fatality rate (26%) that is associated with brainstem encephalitis.9–11 Rapid clinical deterioration and death occurred in the previous outbreaks.12–15
In Taiwan, a small cluster of cases that resulted from EV 71 infection as well as sporadic cases of EV 71 infection occurred in 1980 and 1986.15,16 No children died during either outbreak. From March to December 1998, 405 severe cases of HFMD/herpangina and 78 deaths were reported to Taiwan's Center for Disease Control (CDC). EV 71 circulated in Taiwan for at least 18 years before 1998.17,18 In addition, sequences of some EV 71 isolates in 1998 showed a high degree (92%) of identity in their VP-1 genomic region to that of the EV 71 strain that was isolated in 1986.19 However, underlying factors of the widespread increase in the scale of EV 71 infection in 1998 remain unknown. Therefore, we analyzed the data reported by active and passive surveillance systems in Taiwan to assess the epidemiology of HFMD/herpangina in Taiwan.
Taiwan has a population of ∼22.7 million. The land area is 36188 km,2 so the population density is 627 per km.2 The majority (95%) of the population live in the western part of Taiwan, which we divided into northern, central, and southern regions. Only 5% live in eastern Taiwan, where medical care and socioeconomic status are classified as underprivileged. There are ∼9400 pediatricians; internists; general practitioners; family physicians; and ear, nose, and throat specialists on the island. The sentinel surveillance system has reported cases to the department of health on a weekly basis since 1990 and has been previously described.20 This department enlists 850 representative sentinel physicians from all 22 cities and countries. Each week, sentinel physicians report suspected cases among their ambulatory patients. Diseases such as influenza-like respiratory tract inflections, acute diarrhea, chickenpox, measles, and pertussis were reported before the 1998 epidemic, although HFMD/herpangina was not. Because of the perception that HFMD/herpangina cases were becoming increasingly prevalent, HFMD/herpangina was included in reports beginning on March 3, 1998. The mean number of physicians who reported between 1998 and 2005 was 800, which is 8.5% of all primary physicians.
Another report system, designed for monitoring severe cases of HFMD/herpangina, was established on May 29, 1998. Patients who were hospitalized for HFMD/herpangina were reported to the CDC, Department of Health, by Taiwan's 23 academic medical centers, 80 regional hospitals, and 435 district hospitals, categorized by the hospital size, care, and teaching capacity.15
Patients with HFMD had vesicular lesions on their hands, feet, mouth, and, frequently, buttocks. Lesions in the mouth were often ulcerated. Herpangina is a vesicular exanthem of the fauces and soft palate, often accompanied by fever, sore throat, and pain on swallowing. Most patients who exhibited these manifestations alone were treated as outpatients. A case was defined as severe by the presence of symptoms/signs of HFMD/herpangina in addition to the occurrence of >1 of the following complications: encephalitis, aseptic meningitis, or acute flaccid paralysis; pulmonary edema/hemorrhage; or myocarditis.15
Encephalitis was characterized by a disturbance in the level of consciousness, such as lethargy, drowsiness, or coma. Aseptic meningitis was characterized by headache, meningeal signs, and mononuclear pleocytosis (>5 × 106 leukocytes per L if the patient was older than 1 month or >25 × 106 leukocytes per L if the patient was a newborn) in addition to a negative bacterial culture. Pulmonary edema/hemorrhage was characterized by respiratory distress; tachypnea; tachycardia; pink frothy sputum; and rapidly progressing, patchy, diffuse pulmonary infiltrates and congestion on a chest film. Cardiopulmonary collapse was defined as the development of hypoxemia and hypotension, despite the administration of inotropic drugs. Acute flaccid paralysis was defined as acute onset of paresis or paralysis of 1 or more skeletal muscle groups, usually of 1 or more limbs. Myocarditis was characterized by evidence of decreased contractility on echocardiography, arrhythmia, an enlarged heart, and elevations in cardiac enzymes that are markers of cardiac damage.
Isolation and Identification of Enteroviruses
Annual virological surveillance data (aggregated by HFMD/herpangina season) for Taiwan were obtained from the Epidemiologic Bulletin Reported by the CDC, Taiwan. Viral laboratories were located in 11 hospitals and in the Taiwan CDC. Specimens consisted of throat swabs, stool, cerebrospinal fluid, and, in rare cases, blood samples. They were from inpatients or outpatients who were suspected of having an enteroviral infection. Most of the patients had HFMD/herpangina, with or without complications.
The methods that were used to identify enteroviruses differed in the various laboratories. Monolayers of Vero, rhabdomyosarcoma, and MRC-5 cells were most commonly used for viral isolation. An immunofluorescence assay was used for identification. Cultures that showed a cytopathic effect that was characteristic of enteroviruses were screened for enteroviruses with the use of an enterovirus screening set (catalog No. 3365; Chemicon International, Temecula, CA), which included pan-enterovirus, coxsackievirus B, echovirus, and poliovirus blends. Coxsackievirus A16 and EV 71 were identified with the use of monoclonal antibodies 3323 and 3324. The former can be used to identify either virus, whereas the latter is specific for EV 71. Identification of EV 71 was confirmed by a neutralization test with a polyclonal rabbit antiserum against EV 71, which was prepared during a previous outbreak in 1986, or a rabbit anti-EV 71 serum. Final enterovirus typing was performed by neutralization testing with the use of polyclonal antiserum (American Type Culture Collection, Rockville, MD).15
Because the high proportion of unknown serotypes could lead to an underestimation of the number of individual enteroviruses, reports with unknown serotypes were excluded from the analysis of serotype distribution.21 The annual incidence of severe cases was calculated by dividing the number of severe cases of HFMD/herpangina for children who were younger than 15 years by the population of children the same age as reported between 1998 and 2005 by Taiwan census data. Annual incidences of severe cases were expressed as the number of severe cases per 100000 children. Annual fatality rates were calculated by dividing the number of deaths that resulted from severe cases of HFMD/herpangina for children who were younger than 15 years by the number of severe cases the same year as reported between 1998 and 2005. Annual fatality rates of severe cases were expressed as the number of deaths per 100 severe cases. All statistical analyses were performed using Stata Statistical Software: Release 8.0.22 We used χ2 test with Yates' correction for categorical data. The accepted level of significance for all analyses was P < .05. Relative risks were calculated by the use of Poisson regression analysis.23
Figure 1 shows the number of cases of HFMD/herpangina reported by sentinel physicians from March 1998 through December 2005. During the 8-year study period, epidemic peaks occurred every year, with the highest number of cases during an epidemic that seemed to occur during the summer season. The peak was reached 1 week earlier in the central region and 1.5 weeks later in the southern region. The first wave encompassed all 4 regions of Taiwan. The second wave was largely limited to the southern region, lasting from the first week of September to the second week of December. It peaked during the first week of October. The number of cases reported varied from year to year (range: 0.8–19.9 cases per doctor per week), with the highest number of cases reported in 1998.
The number of cases of severe HFMD/herpangina is shown in Fig 2. With the exception of 1999, peak incidence occurred in the summer season of the year, around the same time as the peak in uncomplicated cases of HFMD/herpangina (Fig 1). A smaller peak is evident in October, representing, as in the case of uncomplicated HFMD/herpangina, cases from southern Taiwan. The smaller number of cases during the second peak was largely restricted to the southern region.
Enterovirus isolate data were available from May 1998 through December 2005. Table 1 shows the data on enteroviruses that were isolated from outpatients and inpatients with HFMD/herpangina. A total of 14910 specimens were tested for enterovirus during the study period with a mean of 1864 specimens (range: 334–2771) each year. An average of 54% (8030 per 14910) of the specimens tested positive for enterovirus. During 1998–2005, coxsackievirus 16 and EV 71 were the predominant serotypes. Each of these serotypes accounted for 23% of reports associated with an identified serotype followed by coxsackievirus B3 (13%), echovirus 4 (6%), coxsackievirus B4 (5%), and echovirus 6 (5%). This demonstrates the various enteroviruses that circulated in the community (Table 1).
From May 1998 through December 2005, there were 900 isolates from inpatients and 7130 isolates from outpatients. The distribution of EV 71, coxsackievirus A16, and other enteroviruses in inpatients and outpatients is shown in Table 2. The relative risk (RR) for an EV 71 isolate was higher among inpatients than outpatients (RR: 16.02; 95% confidence interval [CI]: 13.38–19.19; P < .0001), whereas the RR for a coxsackievirus A16 infection was lower among inpatients than outpatients (RR: 0.58; 95% CI: 0.47–0.71; P < .0001).
Between 1998 and 2005, a total of 1548 severe cases of HFMD/herpangina were reported to the Taiwan CDC. The mean age of patients was 2.2 years (range: 3 months to 14 years), and the male-to-female ratio was 1.5:1.
The annual incidence of severe cases changed significantly during the 8-year study period for both boys and girls (χ2 for linear trend = 165.62; P < .0001). Annual incidences were 9.02 in 1998, 0.85 in 1999, 7.37 in 2000, 9.59 in 2001, 4.10 in 2002, 1.82 in 2003, 1.30 in 2004, and 4.21 in 2005 for boys (χ2 for linear trend = 83.82; P < .0001) and 7.11 in 1998, 0.61 in 1999, 4.86 in 2000, 7.00 in 2001, 2.75 in 2002, 1.20 in 2003, 0.94 in 2004, and 2.27 in 2005 for girls (χ2 for linear trend = 84.93; P < .0001; Fig 3A). We used 1998 as the HFMD/herpangina epidemic year and the remaining 7 years as the baseline years. Compared with baseline years, 1998 had 2.48 times the annual incidence of HFMD/herpangina (95% CI: 2.21–2.78; P < .0001).
Age-specific annual incidence of severe cases by year is shown in Fig 3C. Overall, incidences in all age groups were W-shaped, with peaks in 1998 and 2001. Annual incidence decreased as age increased, demonstrating a peak incidence rate in 1-year-old children (P < .0001).
We plotted the cumulative distribution of cases of severe HFMD/herpangina by age (Fig 4). Most (93%) of the severe cases of HFMD/herpangina occurred in children who were 4 years or younger, with 75% occurring in children who were ≤2 years of age.
Figure 3E shows incidence rates in the 4 regions of Taiwan. The central regions had the highest incidence rate from 1998 through 2005. The lowest incidence rate was observed in the eastern region.
Annual fatality rates changed significantly during the 8-year study period. The annual fatality rates were 20.6 (48 of 233) in 1998, 23.8 (5 of 21) in 1999, 9.9 (18 of 181) in 2000, 15.3 (36 of 235) in 2001, 21.0 (21 of 100) in 2002, 4.7 (2 of 43) in 2003, 16.7 (5 of 30) in 2004, and 10.5 (10 of 95) in 2005 (χ2 for trend = 3.340; P = .068) for boys and 17.4 (30 of 172) in 1998, 28.6 (4 of 14) in 1999, 20.9 (23 of 110) in 2000, 13.9 (22 of 158) in 2001, 14.5 (9 of 62) in 2002, 22.2 (6 of 27) in 2003, 0.0 (0 of 20) in 2004, and 12.8 (6 of 47) in 2005 (χ2 for trend = 1.465; P = .226) for girls (Fig 3B).
The age-specific annual fatality pattern was the difference within different age groups (Fig 3D). Overall, the fatality of HFMD/herpangina was higher among younger than older children, with children who were younger than 1 year having the highest rate during the studied time period (P < .0001).
Fatality rate according to region is shown in Fig 3F. The fatality pattern was different among the 4 regions studied between 1998 and 2005.
Table 3 shows the clinical complications and types of viruses that were isolated from 900 patients with severe infections. Only 1 virus was isolated from each patient. Each complication or combination of complications shown in Table 3 is mutually exclusive. The overall frequency of EV 71 isolates in these patients was 66% (594 of 900). EV 71 was more frequently isolated from patients with encephalitis and pulmonary edema/hemorrhage (93%). Enteroviruses other than EV 71 were more frequently isolated when the complication was aseptic meningitis (65%). Compared with the frequency of EV 71 isolated from patients with aseptic meningitis, a significantly increased frequency of EV 71 was isolated from patients with pulmonary edema/hemorrhage only (RR: 5.74; 95% CI: 2.99–11.09; P < .001) and encephalitis only (RR: 3.14; 95% CI: 2.00–4.94; P < .001).
These data show that an epidemic of HFMD/herpangina that occurred in 1998 is the largest recognized epidemic in Taiwan to date. Various enteroviruses, including EV 71, which is characterized by neurotropism and may cause severe disease or sudden death, circulate predominately in Taiwan each year. Male children and younger children (<5 years) are at an increased risk for infection.
Most patients with HFMD/herpangina are very young children (<4 years), with peak incidence occurring at 1 year of age.24–26 In the United States, 56% of reported enteroviral diseases occurred in children who were younger than 10 years and 26% were younger than 1 year.27 In Belgium, >80% of enteroviral infections were in children who younger than 4 years.28 Our study consistently showed that >93% of severe HFMD/herpangina was found in children who were ≤4 years of age. This may explain why epidemics of HFMD/herpangina occur every 2 to 3 years, because such an interval allows a new population of susceptible individuals to accumulate.29
This study found that male patients outnumbered female patients by 1.5:1. This predominance has been observed in other enteroviral infections, in which the male-to-female ratio ranged from 1.5:1 to 2.5:1.26,30 The reason for this finding is unclear but may suggest a susceptibility at the host genetic level.
Enteroviruses are distributed worldwide. The prevalence pattern of nonpolio enterovirus varies with time and country.28,31–33 In temperate climates, enteroviruses cause epidemics during the summer and autumn months, whereas in tropical areas, infections occur with high incidence throughout the year.32 Some studies show the cyclic recurrence of several virus types.34 Taiwan lies in a temperate climate at a latitude of 21° to 25° north and has a pattern of seasonal enteroviral infection similar to that found in other temperate regions.
This study found that the rural eastern region had the lowest incidences of severe cases. The underlying cause of this remains unknown. Less personal contact as a result of low population density (104 per km2) could be an explanation.
EV 71 can result in disability and even death. The underlying factors of the pathogenesis of pulmonary edema/hemorrhage that is caused by EV 71 infection are unknown. The mysterious hallmark symptom of pulmonary edema/hemorrhage, which can kill a child within 1 day or require hospital admission, may be caused by increased pulmonary vascular permeability as a result of either brainstem encephalitis or a systemic inflammatory response caused by excessive cytokine release.35,36 Another theory suggests that pulmonary edema/hemorrhage is attributable to viral strain differences.15 Either 1 of these 2 theories may be relevant to understanding EV 71 pulmonary syndrome.
There has been a great deal of interest in trying to explain the spread and virulence of EV 71 in molecular and genetic terms. From the genotype analysis of EV 71 collected from isolates in Japan, Malaysia in 1997, and Taiwan in 1998, a group of genetically similar strains have, for the first time, recently been observed to be circulating throughout the western Pacific region,37 although not elsewhere. These strains are different from the EV 71 genotypes that caused previous outbreaks, such as the large epidemics in Bulgaria and Hungary in 1975 and 1978.14,38 Is it the evolution of a more virulent swarm of EV 71, or does the capacity to evolve constantly new and different genotypes represent the competitive edge of this virus?
As with other enteroviruses, the V1, V2, and V3 regions of EV 71 are responsible for the antigenic diversity of enteroviruses, but neutralization epitopes are densely clustered in V1.39 Specific mutations in the 5′-noncoding region of poliovirus have been correlated with neurovirulence.40 Zheng et al41 isolated a strain of EV 71 from mainland China that differed from the prototype BrCr strain in the nucleotide sequence of the 5′-noncoding region. Such differences were thought to determine clinical variability. However, nucleotide sequence analysis of different regions of the EV 71 genome reveals heterogeneity of isolates not only geographically but also temporally.19 No clear marker of neurovirulence or any other clinical manifestation has yet been identified, and we still do not have enough evidence to make any conclusions from this study.
Co-infection with a second virus has been suggested to be another possible pathogenetic factor.15 This theory is supported by the concomitant isolation of a subgenus B adenovirus with enterovirus from 3 people who died during an HFMD outbreak in Malaysia.14 Dual enteroviruses co-circulating in Taiwan were found during several different years throughout the study period. However, the presence of dual viruses did not result in increased incidence and mortality of HFMD/herpangina. Consistent with previous studies,15,19 this study showed that, in addition to EV 71, coxsackievirus A16 and, perhaps, other enteroviruses were circulating in the community, but EV 71 was more prevalent among severely ill and hospitalized patients.
According to the findings of this study, severe cases caused by EV 71 probably had occurred but remained unrecognized. It is not strictly correct to call the Taiwan epidemic an EV 71 epidemic, because large numbers of other enteroviruses, especially coxsackievirus A16, co-circulated, and their infections were a major contributor to HFMD/herpangina. Enterovirus outbreaks are common, although they are newsworthy only when death or disability results. The large number of outbreaks in 1998 could, in part, have been the result of the HFMD/herpangina surveillance initiated in 1998. In addition, physicians, parents, and children's caregivers panicked over the disease as a result of media publicity over the local HFMD/herpangina-related deaths in March 1998. Parents and caregivers sought medical attention for many children, including those with mild illness. Another contributing factor was compulsory reporting of the disease beginning in May 1998.
The data that were used for this study were obtained by a surveillance-based method. The limitation of this study was reporting or selective bias. However, such information is a reasonable surrogate for population level data.
HFMD/herpangina is a common disease in Taiwan. EV 71 infection has emerged as an important public problem causing serious neurologic manifestations and pulmonary edema/hemorrhage and death of young children. Vaccine development is recommended for prevention of EV 71 infection in the future.
The funding for this study was supported by the Taiwanese government.
We thank the staff of the CDC, Taiwan, for assistance in collecting the epidemiologic data. We also thank all staff in the local health bureau and reference laboratories for help in the collection and analysis of specimens.
- Accepted January 26, 2007.
- Address correspondence to Kow-Tong Chen, MD, PhD, Department of Public Health, College of Medicine, National Cheng-Kung University, Taiwan No. 1, University Road, Tainan, Taiwan. E-mail:
The authors have indicated they have no financial relationships relevant to this article to disclose.
- ↵Melnick JL. Enterovirus: polioviruses, coxsackieviruses, echoviruses, and newer enteroviruses. In: Fields BN, Knipe DM, Howley PM, et al, eds. Field's Virology. 3rd ed. Philadelphia, PA: Lippincott-Raven Publishers; 1996: 655‐712
- ↵Tunnessen WW Jr. Erythema infectiosum, roseola, and enteroviral exanthems. In: Gorbach SL, Bartlett JG, Blacklow NR, eds. Infectious Diseases. Philadelphia, PA: WB Saunders Co; 1992: 1120–1125
- ↵Cohen JI. Enterovirus and reovirus. In: Faici AS, Braunwald E, Issebacher KJ, et al, eds. Harrison's Principles of Internal Medicine. 14th ed. New York, NY: McGraw-Hill; 1998: 1118–1123
- ↵Hand, foot, mouth disease in Hong Kong (revised August 11, 1998). Ottawa: Laboratory Center for Disease Control, Health Protection Branch, Health Canada. Available at: www.phac-aspc.gc.ca/tmp-pmv/1998/197_e.html. Accessed November 4, 2006
- ↵Schmidt NJ, Lennette EH, Ho HH. An apparently new enterovirus isolated from patients with disease of the central nervous system. J Infect Dis.1974;129 :304– 309
- ↵Deibel R, Gross LL, Collins DN. Isolation of a new enterovirus. Proc Soc Exp Biol Med.1975;148 :203– 207
- Ishimaru Y, Nakano S, Yamaoka K, Takami S. Outbreaks of hand, foot, and mouth disease by enterovirus 71. High incidence of complication disorders of central nervous system. Arch Dis Child.1980;55 :583– 588
- ↵Wang SM, Liu CC, Tseng HW, et al. Clinical spectrum of enterovirus 71 infection of children in southern Taiwan, with emphasis on the neurological complications. Clin Infect Dis.1999;29 :184– 190
- ↵Kitamura A, Narisawa T, Hayashi A, et al. Serotype determination of enteroviruses that cause hand-foot-mouth disease: identification of enterovirus 71 and coxsackievirus A16 from clinical specimens by using specific probe. J Jpn Assoc Infect Dis.1999;71 :715– 723
- ↵Dushoff J, Plotkin JB, Viboud C, Earn DJD, Simonsen L. Mortality due to influenza in the United States: an annualized regression approach using multiple-cause mortality data. Am J Epidemiol.2006;163 :181– 187
- ↵Stata, StataCorp. Stata Statistical Software [computer program]. Release 8.0. College Station, TX: Stata Corp; 2003
- ↵Rothman KJ. Modern Epidemiology. Philadelphia, PA: Lippincott-Raven; 1998
- ↵Chan LG, Parashar UD, Lye MS, et al. Deaths of children during an outbreak of hand, foot and mouth disease in Sarawak, Malaysia: clinical and pathological characteristics of the disease. For the Outbreak Study Group. Clin Infect Dis.2000;31 :678– 683
- ↵Moore M. Enteroviral disease in the United States, 1970–1979. J Infect Dis.1982;146 :103– 108
- ↵Morens DM, Pallansch MA. Epidemiology. In: Rotbart HA, ed. Human Enterovirus Infections. Washington, DC: American Society for Microbiology; 1995:3–23
- ↵Strikas RA, Anderson LJ, Parker RA. Temporal and geographic patterns of isolates of nonpolio enterovirus in the United States, 1970–1983. J Infect Dis.1986;153 :346– 351
- ↵Bahri O, Rezig D, Nejma-Queslati BB, et al. Enteroviruses in Tunisia: virological surveillance over 12 years (1992–2003). J Med Microbiol.2005;54 :63– 69
- ↵Chang LY, King CC, Hsu KH, et al. Risk factors of Enterovirus 71 infection and associated hand, foot, and mouth disease/herpangina in children during an epidemic in Taiwan. Pediatrics.2002;109(6) . Available at: www.pediatrics.org/cgi/content/full/109/6/e88
- ↵Pallansch MA, Roos RP. Enteroviruses: polioviruses, coxsackieviruses, echoviruses, and newer enteroviruses. In: Knipe DM, Howley PM, eds. Fields Virology. 4th ed. Philadelphia, PA: Lippincott Williams & Wilkins; 2001:723–776
- Copyright © 2007 by the American Academy of Pediatrics