The Risk Factors of Acquiring Severe Hand, Foot, and Mouth Disease: A Meta-Analysis

Objectives The incidence of severe hand, foot, and mouth disease (HFMD) is not low, especially in mainland China in almost every year recently. In this study, we conducted a meta-analysis to generate large-scale evidence on the risk factors of severe HFMD to provide suggestions on prevention and controlling. Methods PubMed, Web of Science, Embase, Cochrane Library, China National Knowledge Infrastructure (CNKI), and Wanfang (Chinese) were searched to identify relevant articles. All analyses were performed using Stata 14.0. Results We conducted a meta-analysis of 11 separate studies. Fever (odds ratio (OR) 7.396, 95% confidence interval (CI) 3.565–15.342), fever for more than 3 days (OR 5.773, 95% CI 4.199–7.939), vomiting (OR 6.023, 95% CI 2.598–13.963), limb trembling (OR 42.348, 95% CI 11.765–152.437), dyspnea (OR 12.869, 95% CI 1.948–85.017), contact with HFMD children (OR 5.326, 95% CI 1.263–22.466), rashes on the hips (OR 1.650, 95% CI 1.303–2.090), pathologic reflexes (OR 3057.064, 95% CI 494.409–19000), Lethargy (OR 31.791, 95% CI 3.369–300.020), convulsions (OR 23.652, 95% CI 1.973–283.592), and EV71 infection (OR 9.056, 95% CI 4.102–19.996) were significantly related to the risk of severe HFMD. We did not find an association between female sex (OR 0.918, 95% CI 0.738–1.142), scatter-lived children (OR 1.347, 95% CI 0.245–7.397), floating population (OR 0.847, 95% CI 0.202–3.549), rash on the hands (OR 0.740, 95% CI 0.292–1.874), rash on the foot (OR 0.905, 95% CI 0.645–1.272), the level of the clinic visited first (below the country level) (OR 5.276, 95% CI 0.781–35.630), breast feeding (OR 0.523, 95% CI 0.167–1.643), and the risk of severe HFMD. Conclusions Fever, fever for more than 3 days, vomiting, limb trembling, dyspnea, contact with HFMD children, rashes on the hips, pathologic reflexes, lethargy, convulsions, and EV71 infection are risk factors for severe HFMD.


Introduction
Hand, foot, and mouth disease (HFMD) is a common childhood infection disease with characteristic features of fever, oral ulcers, and vesicular rashes on the hands, feet, and buttocks. It is caused by a group of enteroviruses, commonly coxsackie A16 and enterovirus-71 (EV-71). e mode of transmission of HFMD is mainly via the fecal-oral route, respiratory droplets, contact with blister fluid of an infected individual, or general close contact with infected individuals. Most HFMD cases were mild and limited to fever and vesicular exanthema on patients' palms, soles, and mouth along with discomfortness at certain levels. However, some severe cases with potentially fatal complications such as brain stem encephalitis (BE) and/or pulmonary edema (PE) show rapid progression that may lead to serious sequelae, even death. In recent years, more and more outbreaks of severe cases have been reported [1,2]. Previous studies have shown that close monitoring and timely intervention may prevent the development of severity and avert the death of severe HFMD [3][4][5].
erefore, it is extremely necessary to identify the risk factors which predict the occurrence of severity. Metaanalysis is a means of increasing the effective sample size under investigation through pooling of data from individual association studies, thus enhancing the statistical power [6]. In order to identify risk factors of acquiring severe HFMD, prevent deterioration, and reduce acute mortality, we conducted this meta-analysis to determine the risk factors for severe HFMD.

Study Selection.
A systematic search of the literature was done in the following electronic databases: PubMed, Web of Science, Embase, Cochrane Library, China National Knowledge Infrastructure (CNKI), and Wanfang (Chinese). e following keywords were used: severe hand foot mouth disease (HFMD), risk factors, and case control study. e quality of the included studies was assessed using the Newcastle-Ottawa Scale, and studies achieving six or more points were considered to be of high quality.

Inclusion and Exclusion
Criteria. Two investigators searched the electronic databases independently according to the following criteria for inclusion: (1) a case-control study including severe and mild disease patient groups; (2) published up to June 2017; (3) diagnosis of severe HFMD and mild HFMD consistent with the criteria defined by us. Abstracts, reviews, case reports, noncomparative studies, and low-quality studies were excluded. In cases of disagreement, a third investigator acted as an arbitrator, and the disagreements were resolved with the research team by discussion.

Data Extraction and Quality Assessment.
e following items were extracted from the included studies: the first author's name, year of publication, source of publication, type of the study, risk factors, total sample size, number of severe and mild HFMD cases, and diagnostic criteria for severe and mild HFMD. e publication bias was evaluated using Egger's test [7]. If P > 0.05, the publication bias exists; otherwise, the publication bias does not exist.

2.4.
Definitions. HFMD cases were divided into two groups according to Guidelines on the Diagnosis and Treatment of HFMD [8]. e mild HFMD was defined as papular/vesicular skin rashes on the hand, foot, mouth, or buttock, and the severe HFMD was defined as mild HFMD with the addition of neurological, respiratory, or circulatory complications, or death. Neurological complications included aseptic meningitis, encephalitis, and acute flaccid paralysis [9]. e duration of fever was defined as body temperature ≥37.5°C.

Meta-Analysis
Methods. Stata 14.0 was used for the statistical analysis. e odds ratio (OR) and 95% confidence intervals (CI) were calculated using the fixed effect model or random effect model, and the choice for statistical model was determined by their heterogeneity which were assessed by the X 2 and I 2 statistics. I 2 >50% and X 2 -statistic (P < 0.1) were considered to show significant heterogeneity, and the random effect model was adopted; otherwise, the fixed effect model was used. e OR and 95% CI were used as summary statistics for the comparison of the following risk factors: fever, fever for more than 3 days, vomiting, limb trembling, dyspnea, contact with HFMD children, rashes on the hips, pathologic reflexes, lethargy, convulsions, EV71 infection, female sex, scatter-lived children, floating population, rash on the hands, rash on the foot, the level of the clinic visited first (below the country level), and breast feeding. e pooled estimate of risk was obtained by the Mantel-Haenszel method in the fixed effect model and by the M-H heterogeneity method in the random effect model. All P values were 2-sided. A P value less than 0.05 was considered to be statistically significant.
e selection process is shown in Figure 1. All the studies were of high quality according to the Newcastle-Ottawa Scale (NOS). e sample sizes of the included studies ranged from 76 to 761 and amounted to 4082 subjects in total. ere were 1640 patients in the severe HFMD group and 2442 patients in the mild HFMD group. e study and patients' characteristics are summarized in Table 1. e two groups were similar with regard to age and gender.

Evaluation of Publication Bias.
Egger's test analysis of total complications was performed. e results are shown in Table 3. Seventeen compared factors had no publication bias; one risk factor did (rashes on the hips).

Discussion
A recent meta-analysis involving 19 separate studies [21] found that clinical characteristics such as duration of fever more than 3 days, body temperature ≥37.5°C, lethargy, vomiting, and EV71 infection were significantly related to the risk of severe HFMD which is consistent with our findings. Other than these risk factors, we also found rashes on the hips or buttocks, limb trembling, dyspnea, pathologic reflexes, convulsions, and contact with HFMD children significantly Full-text articles retrieved for more detailed evaluation (n = 21) Full-text articles excluded (8 did not meet quality criteria, 2 did not have available date) Full-text articles assessed for eligibility (n = 11) Separate studies included in meta-analysis (n = 19)   [20] 2016 China 36 40 * * * * * * 41/39; P � NS 47/48; P > 0.05 NR: not reported; NS: not significant.Now the Newcastle-Ottawa scale is mainly applied in the evaluation of case-control study. e literature was graded in terms of selection, comparability, and outcome, and each aspect consists of a number of assessment items. When the items are up to the requirements, one star can be obtained, of which the comparability can reach a maximum of 2. Six stars ( * * * * * * ) and more were considered to be of high quality.
increased the risk of severe HFMD. However, we found no significant association between rash on the palm, rash on the soles, female gender, scatter-lived children, floating population, the level of the clinic visited first (below the country level), breast feeding, and severe HFMD. Previous studies on gender have different conclusions: Pan et al. [12] thought that female have a lower risk of attacking severe HFMD than male; however, Wang et al. [10], Yang et al. [11], and Liu et al. [13] do not think that there is a connection, and our analysis found that there is no association between gender and severe HFMD.
Both male and female have the same opportunities to develop severe HFMD. Previous studies have also drawn different conclusions regarding scatter-lived children and floating populations: Zhang et al. [15] thought that scatterlived children reduce the risk of acquiring severe HFMD, but Wang et al. [10], Yang et al. [11], and Pan et al. [12] got the opposite conclusion. Zhang et al. [15] thought that floating Li [20] Deng et al. [16] Pan et al. [18] Liu et al. [13] Yang et al. [11] Wang et al. [10] Zhang et al. [15] Study ID 88.9 0.0112 1 children have a lower risk of acquiring severe HFMD; however, Zeng et al. [14] found a higher risk between them. Yang et al. [11] and Pan et al. [12] did not find any association. Our analysis found both scatter-lived children and floating populations is not related to the risk of severe HFMD. Oral ulcers and vesicular rashes on the hands, feet, and hip/buttocks are common signs of HFMD; however, rashes on different parts can lead to different levels of severity of HFMD. Previous studies regarding rashes on the hands and rashes on the hips have different conclusions: Yang et al. [11] thought that rashes on the hands are protective factors of HFMD, and other two studies found no association between them; Zhang et al. [15] thought that rashes on the hips are risk factors of HFMD, and other two studies found no Overall (I-squared = 54.5%, p = 0.111) Li et al. [17] Yang et al. [   Zhang et al. [15] Pan et al. [18] Overall (I-squared = 81.4%, p = 0.005) Note: weights are from random effects analysis 1 9.25  Zhang et al. [15] Pan et al. [18] Overall (I-squared = 0.0%, p = 0.904) 1 2.34  Zhang et al. [15] Pan et al. [18] Overall (I-squared = 7.2%, p = 0.340) 1 2.64

Study ID
Wang et al. [10] Liu et al. [13] Zhang et al. [15] Deng et al. [16] Li et al. [17] Overall (I-squared = 96.2%, p = 0.000)  Liu et al. [13] Pan et al. [12] Wang et al. [10] 3.04 0.329 1 Figure 13: Forest plots showing the results of the meta-analysis regarding female gender. Zhang et al. [15] Pan et al. [12] Wang et al. [10] Study ID 24.8 0.0403 1 Figure 14: Forest plots showing the results of the meta-analysis regarding scatter-lived children. association between them. Our studies confirm previous conclusions about the relationship between rashes on the feet and severe HFMD but found there is no association between rashes on the hands and severe HFMD and also found that rashes on the hips are risk factors. ese remind us to pay close attention to those kinds of patients whose rashes on the hips.
Most of previous studies included in our meta-analysis found that fever, vomiting, lethargy, convulsions, and contact with HFMD children are risk factors of severe HFMD, and our study confirmed it. Our study also confirmed that severe HFMD is associated with fever for more than 3 days, limb trembling/shaking, dyspnea, pathologic reflexes, and EV71 infection again. Any of these factors increase the possibility of developing severe HFMD. erefore, early recognition and meticulous management of patients with these risk factors are required [22,23].
EV71 invades the central nervous system causing severe disease ranging from meningitis to fatal encephalitis [24]. In our study, we retrieved 7 studies that analyzed the association between EV71 infection and severe HFMD, and the meta-analysis showed that EV71 infection was significantly associated with the development of severe HFMD. Enterovirus 71 (EV71) is the key pathogen of HFMD, accounting for 70% severe HFMD cases and 90% HFMDrelated deaths [25]  Zhang et al. [15] Zeng et al. [14] Yang et al. [11] Study ID 10 0.0997 1 Figure 15: Forest plots showing the results of the meta-analysis regarding floating population.
Liu et al. [13] reported that the level of the clinic visited first (below the country level) has a marked impact on severe HFMD development. However, our study did not come to the same conclusion. In addition, Zhang et al. [15] found breast feeding is a protective factor. However, our study did not produce the same results also. ere may be less to do with the number of studies being included and the less number of samples.
erefore, further studies are needed to be conducted to confirm the association between the level of the clinic visited first and severe HFMD as well as the association between breast feeding and severe HFMD.

Conclusions
In conclusion, we found that eleven factors are associated with the severity of HFMD. Previous conclusion regarding the association between fever (body temperature ≥37.5°C), fever more than 3 days, lethargy, vomiting, and EV71 infection and severe HFMD was consistent with our findings. Also, we found rashes on the hips or buttocks, limb trembling/shaking, dyspnea/breathlessness, pathologic reflexes, convulsions/twitch, and contact with HFMD children significantly increased the risk of severe HFMD. But, we found no significant association between rashes on the palm, rashes on the soles, female gender, scatter-lived children, floating population/migrant, the level of the clinic visited first (below the country level), breast feeding, and severe HFMD. Further studies are needed to confirm our findings.

Conflicts of Interest
e authors declare that they have no conflicts of interest.