Factors associated with severe neurologic complications in patients with either hand-foot-mouth disease or herpangina: A nationwide observational study in South Korea, 2009-2014

Background In 2009, a nationwide sentinel surveillance for hand-foot-mouth disease (HFMD) and herpangina (HA) with neurologic complications was initiated in South Korea. We used this surveillance system to investigate the clinical characteristics of patients with either HFMD or HA with neurologic complications, with the aim of determining risk factors for severe neurologic complications. Methods A retrospective review of medical records was conducted on all cases of HFMD and HA with neurologic complications that were reported in the national system between April 1, 2009 and December 31, 2014. A severe case was defined as having HFMD or HA with encephalitis, polio-like syndrome, or cardiopulmonary failure, and less-severe cases were defined as having HFMD or HA with aseptic meningitis. Results A total of 138 cases (less-severe: 90/138, 65.2%; severe: 48/138, 24.8%) were included from 28 hospitals; 28 ineligible cases were excluded. Of 48 severe cases, 27 (56.2%) had encephalitis; 14 (29.2%) had polio-like syndrome; and seven (14.6%) had cardiopulmonary syndrome. The median patient age was 36 months (IQR: 18–60) and 63 (45.7%) patients were female. Most patients completely recovered, except for seven cases that were fatal or resulted in long-term symptoms (5.1%, 3 patients with neurologic sequelae and 4 deaths). In a multivariable logistic regression analysis, lethargy (OR = 4.67, 95% CI: 1.37–15.96, P = 0.014), female sex (OR = 3.51, 95% CI: 1.17–10.50, P = 0.025), and enterovirus A71 (OR = 3.55, 95% CI: 1.09–11.57, P = 0.035) were significantly associated with severe neurologic complications in HFMD and HA patients. Conclusion In patients with HFMD and HA, lethargy, female, and enterovirus A71 may predict severe neurologic complications.


Introduction
Hand-foot-mouth disease (HFMD) is a common viral infection syndrome in children that is characterized by macular, maculopapular, or vesicular rash on the hands, feet, mouth, buttocks, and other possible locations. Herpangina (HA) is characterized by fever and painful vesiculo-ulcerative oral lesions, without an accompanying skin rash [1]. Both HFMD and HA are caused by various enterovirus serotypes, with coxsackievirus A16 and enterovirus A71 (EV-A71) being the most common causative viral strains [1].
Although the majority of patients with HFMD or HA show a mild and self-limiting clinical course, some cases caused by EV-A71 tend to be more severe and even fatal. The cases caused by EV-A71 have relatively high percentage of neurological involvement throughout the course of illness, such as meningitis, encephalitis, and neurogenic pulmonary edema or hemorrhage [2].
In the late 1990s, EV-A71-associated HFMD and HA outbreaks occurred in several countries in the Asia-Pacific region, including Malaysia, Taiwan, and China; these outbreaks resulted in numerous deaths due to severe complications [3][4][5]. Several years later, South Korea also experienced an EV-A71-associated HFMD and HA epidemic, with multiple deaths occurring in 2009 [6][7]. Accordingly, the need for controlling neurologically-involved HFMD or HA became apparent, and the Korean Centers for Disease Control and Prevention (KCDC) designated both as national notifiable infectious diseases.
A previous study in Korea reported that headache and the presence of neurologic signs were significant risk factors for neurologic complications in HFMD or HA [1]. However, most aseptic meningitis usually presents with a mild and reversible clinical course, while encephalitis, polio-like syndrome, and/or cardiopulmonary failure can result in grave consequences, such as permanent neurological sequelae or death [8]. Therefore, investigations regarding factors associated with severe neurologic involvement are necessary. Thus, the aim of this study was to determine risk factors for severe neurologic complications in patients with either HFMD or HA, using data from Korea's National Sentinel Surveillance System for HFMD with neurologic complications (NSSSH).

Study setting
Both HFMD and HA with neurologic complications were designated as national notifiable infectious diseases in 2009, and the NSSSH was established in the same year. All suspected cases diagnosed in tertiary care hospitals (a total of 44 tertiary care hospitals located throughout South Korea) are mandatorily reported by physicians to the KCDC through this surveillance system and cases were reported from 28 hospitals during 2009-2014. We retrospectively reviewed the medical records of all cases reported to this system between April 1, 2009 and December 31, 2014, gathering the following patient data: i) demographic data, ii) perinatal history, iii) clinical features and outcomes, iv) initial laboratory findings, and v) findings on magnetic resonance imaging (MRI). The retrospective review was performed by two investigators in KCDC from November 25, 2014 to February 17, 2015. The total number of patients with HFMD and HA was collected from the Healthcare Big Data Hub of Health Insurance Review & Assessment Service; data were available from 2010-2014 [9].
To examine virology, we correlated our data with the data of the Nationwide Surveillance System for Infection with Enterovirus; this is another nationwide surveillance system. Via this system, the sentinel hospitals voluntarily sent specimens collected from patients with highly suspicious symptoms related to enterovirus infection including HFMD and HA to KCDC. The specimens were collected without consideration of the date of symptom onset. Once KCDC received samples, reverse transcription-PCR (RT-PCR) using TaqMan technology was applied to all stool specimens, throat swabs, and CSF samples in order to detect the enteroviral genome. Detailed information of this process has been described in a previous report [7]. For cases without having the viral test result at the Nationwide Surveillance System for Infection with Enterovirus, we collected enterovirus PCR results conducted at each hospital.
All collected data were anonymized before analysis.

Case definitions
HFMD was defined as presence of vesiculopapular rashes in at least two classical anatomical sites, namely the hands, feet, buttocks, or mouth. HA was defined as presence of mouth ulcers without any skin rash. Aseptic meningitis was defined as the presence of pleocytosis (leukocyte >5 cell/μL) in cerebrospinal fluid (CSF) analysis, with a negative CSF bacterial culture result. Encephalitis was defined as pleocytosis (leukocyte >5 cell/μL) in CSF analysis, with a negative CSF bacterial culture result, and the presence of an altered level of consciousness and/or other neurologic signs. Polio-like syndrome was characterized by the acute onset of areflexic limb weakness and/or decreased deep tendon reflex. Cardiopulmonary syndrome was defined as the presence of respiratory distress, pulmonary edema, and pulmonary congestion on chest radiograph or symptoms and/or signs compatible with myocarditis. We defined "less-severe" cases as cases of HFMD or HA with aseptic meningitis; "severe" cases had HFMD or HA with at least one of the following: encephalitis, polio-like syndrome, or cardiopulmonary syndrome. We excluded any cases that did not meet the case definition of either less-severe or severe.

Statistical analysis
All statistical analyses were performed using SPSS version 21.0 for Windows (IBM Corporation, Armonk, NY, USA). Categorical variables were analyzed by the chi-square test or Fisher's exact test. Continuous variables were analyzed using the Mann-Whitney U test or the independent t-test. A multivariable logistic regression analysis was performed to evaluate the effect of independent variables on risk. In order to avoid collinearity among the variables, we adopted a stepwise regression method with backward elimination. Using a two-tailed test, a Pvalue of < 0.05 was considered to be statistically significant.

Ethics statement
The study protocol was approved by the institutional review boards of KCDC (IRB number: 2014-12EXP-02-P-E), and the requirement for written informed consent from patients was waived.

Demographic characteristics and clinical manifestations
A total of 138 cases were included from 28 tertiary care hospitals (520-2,496 beds) throughout the Korean peninsula (Fig 1), and data of 28 patients were excluded: 11 cases did not meet the diagnostic criteria of HFMD or HA, and 17 cases did not meet the case definition of severe or less-severe. Of the 138 included cases, 128 (92.8%) had HFMD and 10 (7.2%) had HA. The majority of cases were classified as less-severe (65.2%, 90/138), while 48 (34.8%) cases were classified as severe. Among the severe cases, 27 (56.2%) patients presented with encephalitis; 14 (29.2%) with polio-like syndrome; and seven (14.6%) with cardiopulmonary syndrome (Fig 2).
The number of patients with either HFMD or HA and neurologic complications increased gradually from 2009 and showed peak incidence in 2011. The annual number of reported cases decreased thereafter and only two cases were reported in 2014. However, the total number of patients with either HFMD or HA regardless of neurologic complication decreased from 2011 to 2012, but slightly increased in 2013 (Fig 3).
In terms of clinical outcomes, median hospitalization duration was longer in severe cases compared with less-severe cases (less severe: 6 days; severe: 11 days; respectively, P < 0.001). Most patients completely recovered, except for seven cases that were fatal or resulted in longterm symptoms (5.1%, 7/138; 3 patients with neurologic sequelae and 4 deaths). All seven cases occurred in the group of severe case; all fatal cases were from the group of severe case and had cardiopulmonary syndrome (57.1%, 4/7). Each of the three patients survived with neurologic sequelae had encephalitis, polio-like syndrome, or cardiopulmonary syndrome, respectively.

Discussion
This is the first report to describe the clinical characteristics of HFMD or HA with neurologic complications using the NSSSH. In this nationwide study, the median age of patients was 36 months, which is relatively higher than that of patients with HFMD or HA in Malaysia (median, 23 months) [10] and China (mean, 2.17 ± 1.19 years) [11], but similar that in a previous study in South Korea (median, 33.5 months) and Japan (mean, 3.47 ± 2.41 years) [12]. We found that aseptic meningitis was the most common neurologic complication of HFMD and HA, throughout all study years. This finding is concordant with previous studies in Asia-Pacific region. The proportion of aseptic meningitis among cases of HFMD or HA with neurologic complications was 63.6% of 88 cases in a Korean study, approximately 80% of 199 cases in a Japanese study, and 84.7% of 333 cases in a Taiwanese study [1,12,13]. Additionally, the common symptoms of HFMD or HA with neurologic complications found in the present report were similar to findings of previous studies: two previous Korean studies reported that fever, vomiting, and headache were common symptoms [1,7]. Similarly, according to a Chinese study, fever, vomiting, and hypersomnia were the most common symptoms for HFMD with neurologic complications [14]. Similar to the findings from a previous study [15], our findings showed that the pons and medulla were the most commonly-affected sites in the central nervous system. EV-A71 has a tissue tropism in the brain stem, cerebellum, and spinal cord anterior horn cell, and clinical manifestations may vary according to the affected site [16]; for example, when EV-A71 invades the anterior horn or root of the spinal cord, limb impairment may present [17]. The genogroups of EV-A71 may influence the risk of neurologic involvement; Ooi et al. emphasized that EV-A71 genogroup B4 was less likely to have neurologic complications compared with genogroup C1 or B5 [10]. In the present study, C4a was dominant genogroup (79.6%, 43/54) and higher proportion of severe case was observed in genogroup C4a compared with non-C4a (51.2%, 22/43; 9.1%, 1/11; respectively, P = 0.012) (S2 File). However, the preference to affected sites according to genogroups could not be analyzed due to lack of adequate sample size. Previous studies have focused on analyzing the risk factors of fatal cases of HFMD and HA. Chong et al. compared seven fatal and 131 non-fatal HFMD cases and found that atypical physical findings, raised WBC count, vomiting, and the absence of mouth ulcers were predictive of a fatal disease progression [18]. A Chinese case-control study reported that EV-A71, convulsion, dyspnea, cyanosis, coolness of extremities, and vomiting were risk factors for death in children with HFMD [19]. In another Chinese observational study, female sex, lightreflex insensitivity, tachycardia, and higher serum lactate level were independent risk factors for fatality among patients with HFMD with neurologic complications [11]. According to Chang et al., hyperglycemia, leukocytosis, and limb weakness were prognostic factors for pulmonary edema after neurologic involvement among EV-A71-associated HFMD patients; additionally, the authors postulated that hyperglycemia might have resulted from autonomic nervous system dysfunction, which regulates blood glucose homeostasis in HFMD or HA with neurologic complications [20]. Pulmonary edema is the most fulminant complication associated with HFMD and HA, and is caused by the invasion of EV-A71 into the medullary vasomotor center [21].
There are a few studies focused on analyzing the risk factors of HFMD or HA with neurologic complications. A meta-analysis found that prolonged duration of fever (!3 days), lethargy, hyperglycemia, vomiting, increased neutrophil count, EV-A71 infection, and younger age were significantly associated with severe HFMD including neurologic complications [22]. Furthermore, when the causative pathogen was confined to EV-A71, young age, fever, vomiting, mouth ulcers, breathlessness, cold limbs, and poor urine output were associated with neurologic complications [10]. Similarly, the results of the present study indicated that lethargy, female sex, and EV-A71 were the primary risk factors for HFMD or HA with severe neurologic complications.
There were some limitations to our study. Firstly, the NSSSH gathers data from tertiary care hospitals and cases diagnosed at other hospital types were not included; therefore, our results may not be generalizable. Secondly, certain number of patients may not have been included since the reporting to NSSSH was dependent on each physician's compliance. Third, the NSSSH was established during the 2009 epidemic. Cases which occurred during the first quarter in 2009 were not included and the annual incidence in 2009 might be underestimated. Finally, severe cases were compared with less-severe cases, not with non-severe cases (HFMD or HA without neurologic complications), which may have led us to less precise comparison. Despite these limitations, the overall data of this study is likely to be a reasonable approximation of the true value, and our findings may represent the national status of HFMD or HA with neurologic complications in South Korea.
In conclusion, lethargy, female, and EV-A71 in patients with HFMD or HA may predict the development of severe neurologic complications and require early recognition and careful management for patients that exhibit these risk factors.