Background / Hand, foot, and mouth disease (HFMD), in particular those associated with enterovirus 71 (EV71), has caused large outbreaks in the Western Pacific and Southeast Asian countries in the past three decades. There is currently no effective chemoprophylaxis or vaccination for HFMD or EV71 infection. Public health strategies rest on good understanding of the epidemiological features of HFMD. The present series of studies examined the epidemiological characteristics of HFMD in Hong Kong, with a view to better understand the disease epidemiology so as to guide public health actions. / Methods and results of individual studies / Study (1)--characterizing the changing epidemiological features identified from various surveillance systems for HFMD / The trend of HFMD activities from 2001-2009 was analyzed using the sentinel surveillance statistics and HFMD outbreaks. The type of institutions reporting HFMD over time, incidences of outbreaks in 18 districts, and age and sex distribution of affected persons of HFMD outbreaks were analyzed. The clinical presentation, hospitalization rate, complication rate and case fatality of outbreaks were examined. The circulating enterovirues each year were determined by laboratory surveillance findings from 2001-2009 and test for morbidity caused by EV71. Seasonal peak was detected from warmer months of May through July but a smaller winter peak was found from October to December since 2006. An increasing trend of more older children aged above 5 years were infected, from 25.4% in 2001 to 33.0% in 2009 (p=0.01, Mantel-Haenszel Chi-square test). Laboratory surveillance detected a cyclical high activity of EV71 in every 3 to 4 years, which was associated with a higher average hospitalization rate among patients of the HFMD outbreaks reported in corresponding year, although it was only marginally significant (p=0.09, linear regression test). / Study (2)--analyzing the characteristics of EV71 epidemic in 2008 / All EV71 cases diagnosed by PHLC from 1998-2008 were analyzed. The complication and case fatality rates, percentage requiring hospitalization, median duration of hospital care, and the likelihood of being associated with an HMFD outbreak in institution in 2008 were compared with the corresponding rates calculated from cases reported from 1998-2007. Phylogenetic tree was constructed by using the neighbour-joining method and the molecular epidemiology of EV71 detected in 2008 was compared with the past years’ trends. Ninety-eight EV71 cases were reported in 2008, highest in the past decade. The annual incidence was 1.4/100 000 in general population, with highest incidence reported in children aged 0-4 years old (27.9/100 000). 11.2% had complications including meningitis or encephalitis (6.1%), pneumonia (3.1%), acute flaccid paralysis (1.0%), and shock (1.0%). There was only one fatal case (CFR: 1.0%) attributed to interstitial pneumonitis. 45.9% had concurrent HFMD outbreaks in their schools or institutions, and six schools required temporarily class suspension for 14 days. Both the complication rate and CFR werenot significantly different from the corresponding rates of the past 10 years (p=0.12 and 1.00 respectively). Phylogenetic analysis found that most cases reported in 2008 were C4 strains, which were the predominant circulating strains in the past ten years. / Study (3)--examining the association between meteorological parameters and HFMD activity / The sentinel consultation rate of HFMD was tested for any association with the meteorological parameters obtained from the Hong Kong Observatory from 2000-2004. Different regression models were examined to find the best model for predicting HFMD consultation rates from 2005-2009. In multivariate regression analysis, model M2 (in which mean temperature, diurnal difference in temperature, relative humidity and wind speed were positively associated with HFMD) was found to have a higher R2 (0.119) than M0 and M1 models with an R2 of 0.079 and 0.062 respectively, indicating that HFMD consultation rates were better explained using meteorological parameters measured 2 weeks earlier. The predicted trend of HFMD consultation rates for 2005 to 2009 matched well with the observed one (Spearman’s rank correlation coefficient=0.276, P=0.000). Sensitivity analysis showed that the estimated HFMD consultation rates were mostly affected by varying the relative humidity and least affected by wind speed. / Study (4)--determining the basic reproduction number of coxsackievirus A16 and enterovirus 71 using mathematical model / The basic reproduction numbers (R0) of EV71 and CoxA16 from laboratory confirmed HFMD outbreaks reported to DH from 2004-2009 were determined using mathematical model. Thirty four outbreaks were analyzed, 27 due to CoxA16 and seven due to EV71. The median R0 of EV71 was 5.48 with an inter-quartile range of 4.206.51 while median R0 of CoxA16 was 2.50 with an inter-quartile range of 1.963.67. In the sensitivity analysis, R0 of EV71 was significantly higher than that of CoxA16 in whole range of incubation periods, p≦0.025. R0was not associated with outbreak setting, size of the institution or number of persons affected. / Study (5)--assessing the impact of SARS and pandemic influenza H1N1 on transmission of HFMD in Hong Kong / I compared the observed HFMD consultation rates and the projected rates, which were constructed using mathematical model, in defined periods of 2003 and 2009 during which territory-wide public health interventions (including school closure) against Severe Acute Respiratory Syndrome (SARS) and pandemic influenza H1N1 were implemented. There was a reduction of 57.2% (95% C.I.:53.0-60.7%) in observed HFMD consultation rates during SARS period in 2003 and a reduction of 26.7% (95% C.I.:19.5-32.7%) during pandemic influenza H1N1 period in 2009. In 2003, the projected rates were still lower than the observed rates beyond week 31 until almost the end of the year. On the contrary, in 2009, the observed HFMD consultation rates became comparable to that of the projected rates in August, before the end of the defined intervention period. / Conclusions / This thesis bridges the knowledge gaps regarding epidemiological characteristics of HFMD. The changing epidemiology of HFMD, including the cyclical high activity of EV71 warrants vigilant surveillance of its activity in order to guide preventive measures. I have demonstrated that climate parameters may help predict HFMD activity, which could assist in explaining the winter peak detected in recent years and issuing early warning in the future. The R0 of EV71 and CoxA16 were first determined in the literature and I found that R0 of EV71 was higher than R0 of CoxA16. The reduction of transmission of HFMD during the SARS and H1N1 periods suggested that public health measures are effective in reducing the transmission of enteroviruses. / Ma, Siu Keung. / Thesis (M.D.)--Chinese University of Hong Kong, 2013. / Includes bibliographical references (leaves 121-149). / Electronic reproduction. Hong Kong : Chinese University of Hong Kong, [2012] System requirements: Adobe Acrobat Reader. Available via World Wide Web. / Table of content --- p.ii / Acknowledgements --- p.iv / Abbreviations --- p.v / Caption for Tables --- p.vi / Caption for Figures --- p.viii / Précis --- p.1 / Chapter PART I: --- LITERATURE REVIEW ON HFMD --- p.5 / Chapter Chapter 1. --- Current Understanding of Epidemiology of HFMD --- p.6 / Chapter 1.1 --- Causative agents and virology --- p.6 / Chapter 1.2 --- Clinical presentation and management --- p.8 / Chapter 1.3 --- Geographical distribution and past epidemics --- p.14 / Chapter 1.4 --- Host susceptibility and molecular determinants of neruovirulence --- p.26 / Chapter 1.5 --- Routes of transmission and transmission dynamics --- p.27 / Chapter 1.6 --- Knowledge gap identified from literature review --- p.29 / Chapter PART II: --- STUDIES ON EPIDEMIOLOGY OF HFMD IN HONG KONG --- p.33 / Chapter Chapter 2. --- Study Objectives and Main data source for analysis --- p.34 / Chapter 2.1 --- Aim and objectives of this thesis --- p.34 / Chapter 2.2 --- Sentinel surveillance system for monitoring HFMD activity --- p.35 / Chapter 2.3 --- Institutional outbreaks of HFMD reported to DH . --- p.37 / Chapter 2.4 --- EV 71 infection reported to Department of Health --- p.37 / Chapter 2.5 --- Laboratory surveillance for monitoring enteroviruses --- p.37 / Chapter Chapter 3. --- Study (1)--Characterizing the changing epidemiological features identified from various surveillance systems for HFMD --- p.38 / Chapter Chapter 4. --- Study (2)--Analyzing the epidemic of enterovirus 71 in 2008 and its public health implication to Hong Kong --- p.52 / Chapter Chapter 5. --- Study (3)--Examining the association between meteorological parameters and HFMD activity --- p.67 / Chapter Chapter 6. --- Study (4)--Determining the basic reproduction number of coxsackievirus A16 and enterovirus 71 using mathematical model --- p.85 / Chapter Chapter 7. --- Study (5)--Impact of SARS and Pandemic Influenza H1N1 on transmission of HFMD in Hong Kong --- p.100 / Chapter Chapter 8. --- Conclusion --- p.111 / List of publications related to this thesis --- p.119 / References --- p.121

Ce travail a analysé tous les cas de HFMD déclarés à Hai Phong pendant l’épidémie de 2011 et 2012 qui a été la plus importante au Vietnam et la première enregistrée dans le nord du pays. Hai Phong a connu la plus forte incidence au Nord Vietnam. C’était donc un bon modèle pour étudier la dynamique de cette maladie sans interférence et reliquats de précédentes épidémies ou de patients immunoadaptés.La première section est consacrée à une revue de la littérature sur EV-A71 et les entérovirus. La seconde section est divisée en trois chapitres, chacun abordant un aspect spécifique du projet.Le premier chapitre aborde la dynamique de la maladie et le rôle des directives officielles pour la gestion de l’épidémie de 2011-2012. Outre les éléments de base, cette étude apporte des résultats sur l’influence des directives HFMD durant l’épidémie, ce qui n’avait pas encore été fait. La publication des directives a conduit à un accroissement du score de sévérité et d’une réduction du délai entre le premier pic de fièvre et l’admission. Cet effet est associé à un accroissement de la sensibilisation qui a conduit à déclarer la plupart des patients avec des symptômes sévères pour assurer de meilleurs traitements et suivis. Le travail décrit dans ce chapitre a aussi démontré que trois vagues avec des caractéristiques différentes et causées par trois virus différents s’étaient succédées. La vague 1 et la vague 3 ont été causées respectivement par EV-A71 et par une combinaison de CV-A6 et CV-A16 alors que la vague 2 a été causée par un virus inconnu. Ce travail est aussi une analyse intégrative incluant une analyse spatiotemporelle. La maladie semble s’être étendue vers l’est en suivant les rivières pour atteindre les des zones plus peuplées à partir desquelles elle s’est répandue par les routes secondaires locales. Etant donné l’âge moyen des patients, environ 2 ans, la source de contamination doit être cherchée chez les adultes asymptomatiques contaminés lors de leurs activités professionnelles et des mobilités locales.Le deuxième chapitre aborde la phylogénie et la distribution spatiotemporelle de EV-A71 dans le nord du Vietnam et apporte un éclairage sur l’évolution et la dynamique de cet entérovirus. La protéine de capside VP1 a été ciblée. La première conclusion de ce chapitre est que l’épidémie de 2011 et 2012 n’a pas été causée par une souche exogène mais par des souches d’EV-A71 déjà présentes au Nord Vietnam. Ceci indique qu’elles peuvent se maintenir à faible niveau, asymptomatique, en stase génomique et avec une structuration géographique. La cause de l’épidémie devrait donc être recherchée dans le tissu socio-économique plutôt que dans une émergence extérieure. Une autre conclusion de ce chapitre est la corrélation observée ente les groupes de variants I/V et phylogénie, pathogénicité et groupe ethnique. Les profils des mutations I/V aux positions 249, 262 et 284 sur la protéine VP1 pourraient jouer un rôle dans la pathogénicité, ce qui est appuyé par la corrélation entre variants I/V et sévérité/ethnicité.Le dernier chapitre aborde la modélisation mathématique d’une maladie multiphases telle que HFMD. Il est essentiel de détecter aussi tôt que possible une nouvelle vague associée à un nouvel agent. Grace à la grande taille de la cohorte disponible pour ce travail (environ 9000 patients), nous avons pu développer un système d’équations différentielles apportant une forte correspondance avec les données observées. Le modèle a confirmé l’existence de trois vagues en 2011-2012, ayant des niveaux de virulence différents. Il permet aussi de caractériser chaque vague, de détecter l’apparition d’une nouvelle vague et d’associer des groupes patients à un tableau clinique.En conclusion, ce travail de thèse a permis de souligner plusieurs éléments clés à aborder de façon coordonnée afin de faciliter une surveillance efficace de l’HFMD au Vietnam. / This work analyzed all HFMD cases reported in Hai Phong in 2011 and 2012 outbreak which was the largest to have ever occurred in Vietnam and the first recorded in the northern part of the country. Hai Phong city experienced the highest HFMD incidence in North Vietnam. It was thus a good model for investigating the dynamic of the disease without interference and potential remains from previous outbreaks or patient immunological adaptation.The first section is dedicated to a review of the literature on EV-A71 and enteroviruses. The second section is divided in three chapters, each one addressing a specific issue of the project.The first chapter addresses the dynamic of the disease and the role of official guidelines in the handling of the 2011-2012 epidemic. Beside basic epidemiological features, the study also provides findings relating to the influence of HFMD guidelines during the outbreak period that has never been described before. The guideline release led to a significant increase of the severity score and reduced delay between onset and admission. This effect is linked to an increased awareness leading to patients being mostly declared with severe symptoms in order to ensure a better treatment and surveillance. The work presented in this chapter also demonstrated that three waves occurred with different characteristics and caused by three different viruses. Wave 1 and wave 3 were caused by EV-A71 and a combination of CV-A6 and CV-A16, respectively while Wave 2 was caused by an unknown virus. This work is also an integrative analysis including a spatiotemporal analysis. The disease seems to have expanded following the eastbound river system to reach densely populated settlements from where it secondarily expanded through local roads. Owing to the average age of the patients, around 2, the source of contamination must be sought for within asymptomatic adults being contaminated during their occupational activities and in local movements.The second chapter addresses the phylogeny and spatiotemporal distribution of EV-A71 in North Vietnam and provides an insight on the evolution and dynamic of the EV-A71 enterovirus. The VP1 capsid protein was used as target. The first conclusion of this chapter is that the 2011-2012 outbreak was not caused by an incoming strain but by EV-A71 strains which were already present in North Vietnam. This indicates that they can remain in a low level, asymptomatic state, in genomic stasis and with a geographic structuration. The cause for outbreaks should thus be sought for in the socio-economic patterns rather than in exogenous emergence. Another outcome of this chapter is the observed correlation between I/V variant groups and phylogeny, pathogenicity and ethnicity. The I/V pattern at positions 249, 262 and 284 on the VP1 protein might play a role in pathogenicity. The observed correlation of the I/V variant populations with severity and ethnicity strengthen this hypothesis.The last chapter addresses the mathematical modelling of a multiphase disease such as HFMD. It is essential to detect as soon as possible the emergence of new wave, associated to a novel agent. Owing to the large size of the cohort available for this work (ca. 9000 patients), we have been able to develop a differential equation model providing a very high fit with the observed data. The model confirmed that three waves were present in 2011-2012 with differing virulence. It also allows to characterize each wave, detect the start of a new one and associate groups of patients with specific patterns of symptoms.As a conclusion, this PhD work as underlined some key issues to be addressed in a coordinated way in order help developing an efficient surveillance and monitoring system for HFMD in Vietnam.

No description available.

La maladie pied-main-bouche (PMB) et l’herpangine sont deux maladies pédiatriques bénignes causées par les entérovirus (EV), en particulier les sérotypes de l’espèce A (EV-A). Le sérotype EV-A71 fait l’objet d’une surveillance dans les pays du Sud Est de l’Asie car il est associé à des atteintes neurologiques sévères chez les très jeunes enfants, parfois mortelles (défaillance cardio-pulmonaire). Les infections causées par les autres EV-A tel que le coxsackievirus A16 (CV-A16) provoquent rarement des atteintes sévères. En Europe, les cas de maladie PMB causés par l’EV-A71 ne font pas l’objet d’une déclaration obligatoire, car ce virus ne cause pas d’épidémies de grande ampleur. L’objectif général de la thèse était d’étudier l’épidémiologie des EV-A en Europe et nous avons utilisé une approche phylogénétique bayésienne pour analyser un échantillon de 500 souches. Nous montrons la circulation discontinue de l’EV-A71 de deux populations virales principales (sous génogroupes C1 et C2), ce qui explique la rareté des épidémies en Europe. L’épidémiologie de ce virus est aussi caractérisée par des transports de souches entre les pays Européens et sporadiquement entre l’Europe et l’Asie (sous génogroupes B5 et C4). La recombinaison génétique intertypique survient rarement parmi les populations d’EV-A71 en circulation et ne contribue pas significativement à leur diversité génétique. Cependant, ce mécanisme génétique est relié à l’émergence d’un sous génogroupe CV-A16 qui circule en France depuis 2011. Comparés à l’EV-A71, les sérotypes CV-A2, CV-A4, CV-A6 sont plus fréquemment sujets à des événements de recombinaison intertypiques. L’analyse de la sélection à l’échelle moléculaire indique que la fixation des mutations dans les protéines de capside de l’EV-A71 est lente, probablement à cause des contraintes structurales et fonctionnelles. La surveillance des infections à EV-A71 en Europe devrait être renforcée à cause de la neurovirulence de ce virus, de l’introduction récente et répétée de souches variantes « asiatiques » et de l’existence d’une grande diversité de génogroupes en Afrique et en Inde encore peu explorée. / Hand-Foot and Mouth Disease (HFMD) and Herpangina are two benign pediatric diseases caused by Enteroviruses (EV), especially enterovirus A species serotypes (EV-A). Infections caused by the EV-A71 serotype are monitored in countries of South East Asia because they are associated with severe neurological symptoms in young children and may be fatal (cardiopulmonary failure). Infections caused by the other EV-A serotypes, e.g. coxsackievirus A16 (CV-A16), rarely induce severe symptoms. In Europe, EV-A71 HFMD cases are not notifiable because this virus does not cause large-scale epidemics. The overall objective of this thesis was to study the EV-A epidemiology in Europe and we used a Bayesian phylogenetic approach to analyze 500 viral strains. We show a discontinued circulation of two EV-A71 populations (C1 and C2 subgenogroups), which explains the rare outbreaks in Europe. The epidemiology of this virus is characterized by transportation events of viral strains between European countries and sporadically between Europe and Asia (C4 and B5 subgenogroups). Intertypic genetic recombination occur rarely among circulating EV-A71 populations and does not contribute significantly to their genetic diversity. We found that genetic mechanism was related to the emergence of a new CV-A16 subgenogroup, which is circulating in France since 2011. In comparison with EV-A71, a number of serotypes (CV-A2, CV-A4, and CV-A6) are more frequently involved in intertypic recombination events. The structural and functional constraints are possible factors involved in the slow mutation fixation in the EV-A71 capsid proteins as determined by analyses of molecular selection. Neurovirulence, the recent and repeated introductions of variants “Asian” strains, and the diversity of genogroups in Africa and India call for strengthened surveillance of EV-A71 infections among European countries.