Background & Objectives: There has been few investigation of the association between the farming related activities or specific characteristics and atopic disease in rural Canadian children. In population-based studies, assuring the quality of information from questionnaires is of concern. We conducted this study in order to: first, identify the prevalence and risk factors of atopy and allergic conditions among school-age children in a rural region of Canada. Also, we sought to evaluate the validity and reliability of a questionnaire report of allergy to assess in this population. Methods: As part of a longitudinal study of lung health in rural residents, we conducted a cross-sectional baseline study in rural Saskatchewan, Canada. This included an initial survey phase followed by a clinical testing phase. A sub-sample of 584 children (grades 1-8) completed skin prick testing to assess atopic status. Of these, 480 children completed a questionnaire report of allergy and atopic outcomes and participated in skin prick testing (SPT). Atopy was defined as a positive reaction to any of 6 allergens (local grasses, wheat dust, cat dander, house dust mite, Alternaria, Clasdosporium)≥3mm compared to the negative control. Agreement between questionnaire report and objective measures of atopy was considered overall and between the specific allergens tested on SPT and those assessed on questionnaire. We considered percent concordance, Kappa, sensitivity, specificity, and the positive and negative predictive values of reported allergies or allergic conditions in comparison to SPT as the gold standard. Results: The prevalence of atopy as well as allergen-specific sensitizations was similar between farm and non-farm children but supported the notion that livestock farming is protective against atopy. Also, we found that 25.0% of children reported a history of allergic conditions by questionnaire and 19.4% were atopic detected by skin pick test. In our study, the agreement between questionnaire report of specific allergic triggers and atopy measured by SPT was high (83.0% - 89.5%). Conclusion: In children, livestock exposure has a protective effect on SPT positivity.The agreement between questionnaire report of allergic symptoms and atopy measured by SPT was high and the agreement between atopy and report of allergic conditions was moderate.

Parkinson’s disease (PD) is a progressive, degenerative, neurological disease. The progressive disability associated with PD results in substantial burdens for those with the condition, their families and society in terms of increased health resource use, earnings loss of affected individuals and family caregivers, poorer quality of life, caregiver burden, disrupted family relationships, decreased social and leisure activities, and deteriorating emotional well-being. Currently, no cure is available and the efficacy of available treatments, such as medication and surgical interventions, decreases with longer duration of the disease. Whilst the cause of PD is unknown, genetic and environmental factors are believed to contribute to its aetiology. Descriptive and analytical epidemiological studies have been conducted in a number of countries in an effort to elucidate the cause, or causes, of PD. Rural residency, farming, well water consumption, pesticide exposure, metals and solvents have been implicated as potential risk factors for PD in some previous epidemiological studies. However, there is substantial disagreement between the results of existing studies. Therefore, the role of environmental exposures in the aetiology of PD remains unclear. The main component of this thesis consists of a case-control study that assessed the contribution of environmental exposures to the risk of developing PD. An existing, previously unanalysed, dataset from a local case-control study was analysed to inform the design of the new case-control study. The analysis results suggested that regular exposure to pesticides and head injury were important risk factors for PD. However, due to the substantial limitations of this existing study, further confirmation of these results was desirable with a more robustly designed epidemiological study. A new exposure measurement instrument (a structured interviewer-delivered questionnaire) was developed for the new case-control study to obtain data on demographic, lifestyle, environmental and medical factors. Prior to its use in the case-control study, the questionnaire was assessed for test-retest repeatability in a series of 32 PD cases and 29 healthy sex-, age- and residential suburb-matched electoral roll controls. High repeatability was demonstrated for lifestyle exposures, such as smoking and coffee/tea consumption (kappas 0.70-1.00). The majority of environmental exposures, including use of pesticides, solvents and exposure to metal dusts and fumes, also showed high repeatability (kappas >0.78). A consecutive series of 163 PD case participants was recruited from a neurology clinic in Brisbane. One hundred and fifty-one (151) control participants were randomly selected from the Australian Commonwealth Electoral Roll and individually matched to the PD cases on age (± 2 years), sex and current residential suburb. Participants ranged in age from 40-89 years (mean age 67 years). Exposure data were collected in face-to-face interviews. Odds ratios and 95% confidence intervals were calculated using conditional logistic regression for matched sets in SAS version 9.1. Consistent with previous studies, ever having been a regular smoker or coffee drinker was inversely associated with PD with dose-response relationships evident for packyears smoked and number of cups of coffee drunk per day. Passive smoking from ever having lived with a smoker or worked in a smoky workplace was also inversely related to PD. Ever having been a regular tea drinker was associated with decreased odds of PD. Hobby gardening was inversely associated with PD. However, use of fungicides in the home garden or occupationally was associated with increased odds of PD. Exposure to welding fumes, cleaning solvents, or thinners occupationally was associated with increased odds of PD. Ever having resided in a rural or remote area was inversely associated with PD. Ever having resided on a farm was only associated with moderately increased odds of PD. Whilst the current study’s results suggest that environmental exposures on their own are only modest contributors to overall PD risk, the possibility that interaction with genetic factors may additively or synergistically increase risk should be considered. The results of this research support the theory that PD has a multifactorial aetiology and that environmental exposures are some of a number of factors to contribute to PD risk. There was also evidence of interaction between some factors (eg smoking and welding) to moderate PD risk.

Parkinson’s disease (PD) is a progressive, degenerative, neurological disease. The progressive disability associated with PD results in substantial burdens for those with the condition, their families and society in terms of increased health resource use, earnings loss of affected individuals and family caregivers, poorer quality of life, caregiver burden, disrupted family relationships, decreased social and leisure activities, and deteriorating emotional well-being. Currently, no cure is available and the efficacy of available treatments, such as medication and surgical interventions, decreases with longer duration of the disease. Whilst the cause of PD is unknown, genetic and environmental factors are believed to contribute to its aetiology. Descriptive and analytical epidemiological studies have been conducted in a number of countries in an effort to elucidate the cause, or causes, of PD. Rural residency, farming, well water consumption, pesticide exposure, metals and solvents have been implicated as potential risk factors for PD in some previous epidemiological studies. However, there is substantial disagreement between the results of existing studies. Therefore, the role of environmental exposures in the aetiology of PD remains unclear. The main component of this thesis consists of a case-control study that assessed the contribution of environmental exposures to the risk of developing PD. An existing, previously unanalysed, dataset from a local case-control study was analysed to inform the design of the new case-control study. The analysis results suggested that regular exposure to pesticides and head injury were important risk factors for PD. However, due to the substantial limitations of this existing study, further confirmation of these results was desirable with a more robustly designed epidemiological study. A new exposure measurement instrument (a structured interviewer-delivered questionnaire) was developed for the new case-control study to obtain data on demographic, lifestyle, environmental and medical factors. Prior to its use in the case-control study, the questionnaire was assessed for test-retest repeatability in a series of 32 PD cases and 29 healthy sex-, age- and residential suburb-matched electoral roll controls. High repeatability was demonstrated for lifestyle exposures, such as smoking and coffee/tea consumption (kappas 0.70-1.00). The majority of environmental exposures, including use of pesticides, solvents and exposure to metal dusts and fumes, also showed high repeatability (kappas >0.78). A consecutive series of 163 PD case participants was recruited from a neurology clinic in Brisbane. One hundred and fifty-one (151) control participants were randomly selected from the Australian Commonwealth Electoral Roll and individually matched to the PD cases on age (± 2 years), sex and current residential suburb. Participants ranged in age from 40-89 years (mean age 67 years). Exposure data were collected in face-to-face interviews. Odds ratios and 95% confidence intervals were calculated using conditional logistic regression for matched sets in SAS version 9.1. Consistent with previous studies, ever having been a regular smoker or coffee drinker was inversely associated with PD with dose-response relationships evident for packyears smoked and number of cups of coffee drunk per day. Passive smoking from ever having lived with a smoker or worked in a smoky workplace was also inversely related to PD. Ever having been a regular tea drinker was associated with decreased odds of PD. Hobby gardening was inversely associated with PD. However, use of fungicides in the home garden or occupationally was associated with increased odds of PD. Exposure to welding fumes, cleaning solvents, or thinners occupationally was associated with increased odds of PD. Ever having resided in a rural or remote area was inversely associated with PD. Ever having resided on a farm was only associated with moderately increased odds of PD. Whilst the current study’s results suggest that environmental exposures on their own are only modest contributors to overall PD risk, the possibility that interaction with genetic factors may additively or synergistically increase risk should be considered. The results of this research support the theory that PD has a multifactorial aetiology and that environmental exposures are some of a number of factors to contribute to PD risk. There was also evidence of interaction between some factors (eg smoking and welding) to moderate PD risk.

Uma característica dominante do sistema financeiro contemporâneo é a intrincada rede de conexões entre instituições financeiras, destacando-se a rede de empréstimos do mercado interbancário, através da qual é feita a transferência de recursos líquidos de bancos com superavit de liquidez para bancos deficitários. Ao mesmo tempo em que o mercado interbancário é responsável pela alocação eficiente de liquidez, a estrutura das exposições interbancárias pode ser considerada fator de risco sistêmico por ser fonte de contágio em caso de crise financeira. A insolvência de um banco pode se propagar na rede levando à insolvência de um grande subconjunto conectado de bancos. Estudos empíricos tem evidenciado que algumas redes interbancárias apresentam características de redes livres de escala. O presente trabalho explora as características de contágio financeiro em redes cuja distribuição de links se aproxima a uma lei de potência, através de um modelo deliberadamente simplificado que define a estrutura patrimonial dos bancos a partir de informações de conectividade da rede. Variando os parâmetros de formação das redes obtemos distribuições com diferentes concentrações de dívidas e de direitos, criando três perfis principais, que foram analisados quanto a sua resistência ao contágio. Testamos também o efeito da variação da conectividade em conjunto com a variação da concentração dos links. Os resultados encontrados sugerem que redes mais conectadas e com alta concentração de direitos (com nodos caracterizados por serem grandes credores do sistema) apresentam maior resistência ao contágio. Avaliando alguns índices topológicos de risco sistêmico sugeridos na literatura, pudemos verificar sua capacidade de explicar o impacto da quebra de um nodo sobre o sistema. Embora fique evidente a relação positiva entre os índices e o valor do impacto para os casos de maior magnitude de perdas, a relação é mais fraca para os menores valores de impacto, sugerindo um poder menor de previsão em redes mais resistentes. / One of the most striking characteristics of modern financial systems is its complex interdependence, standing out the network of bilateral exposures in interbank market, through which institutions with surplus liquidity can lend to those with liquidity shortage. While the interbank market is responsible for efficient liquidity allocation, it also introduce the possibility for systemic risk via financial contagion. Insolvency of one bank can propagate through interlinkages leading to insolvency of other banks. Empirical studies have shown that some interbank networks have features of scalefree networks. This work explores the characteristics of financial contagion in networks whose links distributions approaches a power law, using a deliberately simplified model that defines banks balance sheets from information of network connectivity. Varying the parameters of the network creation we obtained links distributions with different concentrations of debts and rights, creating three main network types, which were analyzed for their resilience to contagion. We also tested the effect of a variation in connectivity in conjunction with variation in concentration of links. The results suggest that more connected networks with high concentration of rights (featuring nodes that are large creditors of the system) present greater resilience to contagion. Evaluating some topological indices of systemic risk suggested in the literature, we could verify its ability to explain the impact on the system caused by the failure of a node. While it is clear the positive relationship between the indexes and the impact value for cases of greater magnitude of losses, the relationship is weaker for smaller values of impact, suggesting a lower predictive power in more resilient networks.

Uma característica dominante do sistema financeiro contemporâneo é a intrincada rede de conexões entre instituições financeiras, destacando-se a rede de empréstimos do mercado interbancário, através da qual é feita a transferência de recursos líquidos de bancos com superavit de liquidez para bancos deficitários. Ao mesmo tempo em que o mercado interbancário é responsável pela alocação eficiente de liquidez, a estrutura das exposições interbancárias pode ser considerada fator de risco sistêmico por ser fonte de contágio em caso de crise financeira. A insolvência de um banco pode se propagar na rede levando à insolvência de um grande subconjunto conectado de bancos. Estudos empíricos tem evidenciado que algumas redes interbancárias apresentam características de redes livres de escala. O presente trabalho explora as características de contágio financeiro em redes cuja distribuição de links se aproxima a uma lei de potência, através de um modelo deliberadamente simplificado que define a estrutura patrimonial dos bancos a partir de informações de conectividade da rede. Variando os parâmetros de formação das redes obtemos distribuições com diferentes concentrações de dívidas e de direitos, criando três perfis principais, que foram analisados quanto a sua resistência ao contágio. Testamos também o efeito da variação da conectividade em conjunto com a variação da concentração dos links. Os resultados encontrados sugerem que redes mais conectadas e com alta concentração de direitos (com nodos caracterizados por serem grandes credores do sistema) apresentam maior resistência ao contágio. Avaliando alguns índices topológicos de risco sistêmico sugeridos na literatura, pudemos verificar sua capacidade de explicar o impacto da quebra de um nodo sobre o sistema. Embora fique evidente a relação positiva entre os índices e o valor do impacto para os casos de maior magnitude de perdas, a relação é mais fraca para os menores valores de impacto, sugerindo um poder menor de previsão em redes mais resistentes. / One of the most striking characteristics of modern financial systems is its complex interdependence, standing out the network of bilateral exposures in interbank market, through which institutions with surplus liquidity can lend to those with liquidity shortage. While the interbank market is responsible for efficient liquidity allocation, it also introduce the possibility for systemic risk via financial contagion. Insolvency of one bank can propagate through interlinkages leading to insolvency of other banks. Empirical studies have shown that some interbank networks have features of scalefree networks. This work explores the characteristics of financial contagion in networks whose links distributions approaches a power law, using a deliberately simplified model that defines banks balance sheets from information of network connectivity. Varying the parameters of the network creation we obtained links distributions with different concentrations of debts and rights, creating three main network types, which were analyzed for their resilience to contagion. We also tested the effect of a variation in connectivity in conjunction with variation in concentration of links. The results suggest that more connected networks with high concentration of rights (featuring nodes that are large creditors of the system) present greater resilience to contagion. Evaluating some topological indices of systemic risk suggested in the literature, we could verify its ability to explain the impact on the system caused by the failure of a node. While it is clear the positive relationship between the indexes and the impact value for cases of greater magnitude of losses, the relationship is weaker for smaller values of impact, suggesting a lower predictive power in more resilient networks.

Svenska livbolag har varit en av de större kapitalplacerarna på den svenska fastighetsmarknaden under de senaste fem åren. Livbolagen förvaltar kapital åt sina försäkringstagare som förväntar sig få utbetalningar antingen genom en traditionell försäkring, där livbolagen utlovar en bestämd avkastning per år eller genom att erbjuda fondförsäkringar. Den traditionella livförsäkringen inbringar stora summor premier som måste investeras i olika tillgångsslag. Placeringstillgångarna kommer att bli underställda det nya Solvens II direktivet som innebär att livbolagen måste anpassa sina tillgångar efter sina åtaganden. Syftet med den här rapporten är att utreda vilka effekter Solvens II direktivet kan tänkas ha på de svenska livbolagen och hur livbolagen kan komma att förändra sina portföljeallokeringar till följd av det nya regelverket. Rapporten avgränsas till livbolagens fastighetsrelaterade tillgångar och kommer primärt behandla effekten av ett riskbaserat kapitalkrav på olika fastighetsexponeringarna. Avslutningsvis diskuteras ifall livbolagens agerande till följd av direktivet kan komma att påverka fastighetsmarknaden. Utredningen kring förslaget har pågått sedan 2006 och implementeringen förväntas ske kring 2016. Dock så har både Europaparlamentet och Finansinspektionen gett indikationer på att implementeringen kan komma att förskjutas ytterligare. Utöver kapitalkravet kommer Solvens II leda till en mer betungande rapportering samt strukturella förändringar inom livbolagens ledning. Uppsatsen kommer fram till att ingen fastighetsinvestering kommer att premieras för samtliga livbolag men att de kan premieras en fastighetsinvestering för enskilda livbolag beroende på vad de har i sin befintliga portfölj och hur investeringen kan komma att påverka diversifieringseffekten. Eftersom livbolagens portföljer är relativt olika kommer Solvens II inte att påverka den svenska fastighetsmarknaden och om den så var fallet skulle detta redan ha skett eftersom direktivet har varit på agendan sedan 2006. / The Swedish life insurance companies have been one of the major capital investors in the Swedish real estate market in the past five years. The life insurance companies manages capital for its policyholders who expect to get payments either through a traditional insurance where the life insurance companies promise a fixed return per year or by providing a unit-linked insurance. The traditional life insurance brings in huge amounts of premiums that must be invested in different assets. The allocation of the assets will be subjected to the new Solvency II directive, which means that life insurance companies must adapt their assets according to their commitments. The purpose of this report is to investigate the effects the Solvency II directive could have on the Swedish life insurance companies and how life insurance companies may change their allocation of their portfolio as result of these new regulations. The report is limited to the life insurance companies' real estate-related assets and will primarily deal with the impact of risk-based capital requirements for various real estate exposures. Finally, we discuss if the life insurance companies' behavior may affect the Swedish real estate market as a result. The investigation about the proposal has been ongoing since 2006 and implementation is expected to occur around 2016. However, the European Parliament and the Swedish Financial Supervisory Authority have given indications that the implementation could be postponed. In addition to capital requirements, Solvency II will lead to a more stringent reporting and structural changes in the life insurance companies' management. The main findings in this report are that none of the real estate asset classes will be preferred for all life insurances companies, however some asset classes may be preferred from the individual companies perspective with regard of what their present portfolio contains and how the inclusion that asset may affect the diversifying effect. Since every life insurance company portfolio contains a different mix of assets the Solvency II directive will not affect the Swedish real estate market and if that were the case it would already have happened because the directive have been on the agenda since 2006.

iv Abstract This diploma thesis examines the impact of the COVID-19 crisis on the bank credit risk in the European Union. The analysis is performed using two sets of panel data. The first set contains data at the bank-level between 2012 and 2018 and is obtained from BankFocus batabase and the second set of data is obtained from the EBA Risk dashboard and contains data at the country-level between 2014 and 2020. Both datasets contain bank-specific variables and macroeconomic variables. We use the variables Cost of risk, Total capital ratio, Tier 1 ratio and NPE ratio as dependent variables. As representatives of the COVID-19 shock, we use the number of people infected with this disease, the number of deaths from this disease and the Stringency Index. We employ the GMM system for our analysis and test 5 hypotheses. We did not reject 3 hypotheses, namely that Cost of risk is a key determinant of credit risk and that the crisis caused by COVID-19 affects the variables Capitalo ratio and NPE ratio. We further concluded that the variables representing COVID-19 do not have a negative effect on credit risk, mainly due to the interventions of the ECB and the IASB. JEL Classification C12, C33, G01, G21 Keywords bank, COVID-19 crisis, credit risk management, Stringency index Title Author's e-mail Supervisor's e-mail...

Indiana University-Purdue University Indianapolis (IUPUI) / Time-to-event outcomes are widely utilized in medical research. Assessing the cumulative effects of time-varying exposures on time-to-event outcomes poses challenges in statistical modeling. First, exposure status, intensity, or duration may vary over time. Second, exposure effects may be delayed over a latent period, a situation that is not considered in traditional survival models. Third, exposures that occur within a time window may cumulatively in uence an outcome. Fourth, such cumulative exposure effects may be non-linear over exposure latent period. Lastly, exposure-outcome dynamics may differ among groups defined by individuals' characteristics. These challenges have not been adequately addressed in current statistical models. The objective of this dissertation is to provide a novel approach to modeling group-specific dynamics between cumulative timevarying exposures and a time-to-event outcome. A framework of group-specific dynamic models is introduced utilizing functional time-dependent cumulative exposures within an etiologically relevant time window. Penalizedspline time-dependent Cox models are proposed to evaluate group-specific outcome-exposure dynamics through the associations of a time-to-event outcome with functional cumulative exposures and group-by-exposure interactions. Model parameter estimation is achieved by penalized partial likelihood. Hypothesis testing for comparison of group-specific exposure effects is performed by Wald type tests. These models are extended to group-specific non-linear exposure intensity-latency-outcome relationship and group-specific interaction effect from multiple exposures. Extensive simulation studies are conducted and demonstrate satisfactory model performances. The proposed methods are applied to the analyses of group-specific associations between antidepressant use and time to coronary artery disease in a depression-screening cohort using data extracted from electronic medical records.

Un certain nombre d'études ont étudié le lien entre cancer du sein et exposition aux dioxines et au cadmium. Toutefois, les résultats de ces études ne sont pas concluants et présentent des limites méthodologiques. L'objectif principal de cette thèse était de développer un indicateur spatial permettant d'estimer les expositions atmosphériques chroniques aux dioxines et au cadmium des femmes issues de la cohorte E3N en France, entre 1990 et 2008. Nous avons créé une base de données de 2620 sources de dioxines et 2700 sources de cadmium et estimé et géolocalisé leurs émissions. L'estimation des émissions à la source et leur localisation ont servi de base pour la construction de l'indicateur spatial. Nous avons identifié une sélection de paramètres permettant d'obtenir un accord « substantiel » entre les classifications des expositions des femmes de la cohorte E3N, localisées à l'adresse de résidence, estimées avec l'indicateur, et avec le modèle gaussien SIRANE. En utilisant le modèle SIRANE, nous avons également étudié la variabilité spatio-temporelle (1990-2008) des concentrations de dioxines et de cadmium sur la métropole de Lyon. Pour la première fois des concentrations modélisées ont été comparées à des mesures de concentrations de dioxines en air ambiant. L'indicateur développé dans cette thèse a été utilisé pour estimer le risque de cancer du sein associé à l'exposition atmosphérique aux dioxines dans une étude cas-témoins niché au sein de la cohorte E3N. Il est utilisé actuellement dans une étude sur le lien entre exposition au cadmium et risque de cancer du sein, et pourra être appliqué dans de futures études portant sur d'autres polluants ou d'autres pathologies / A number of studies have examined the link between breast cancer and exposure to air pollution, including dioxins and cadmium. However, the results of these studies are inconclusive and present a number of methodological limitations. The main objective of this thesis was to develop a spatial indicator to assess chronic atmospheric exposure to dioxins and cadmium of women from the E3N cohort in France between 1990 and 2008. We first performed an inventory and created a database of 2620 dioxins and 2700 cadmium emitting sources in France between 1990 and 2008 and estimated and geolocated their annual emissions. The location of the sources and their estimated emissions were used as the basis for the construction of the indicator. Combination of additional spatial parameters, allowed us to obtain a "substantial" agreement between the dioxin and cadmium exposure classifications of the E3N subjects geolocalised at their residential address, using the estimated bythe indicator and exposures estimates derived from the Gaussian model. We also evaluated the spatial-temporal variability of dioxin and cadmium concentrations over nearly two decades in the Lyon metropolitan area, taking into account a wide variety of source types. This is the first study comparing concentrations predicted by a dispersion model to dioxin concentrations measured in ambient air. The exposure indicator was used in a case-control study within the E3N cohort to estimate the risk of breast cancer associated with atmospheric exposure to dioxins. It is currently being used in another study on cadmium exposure and breast cancer risk and may be applied in future studies on other pollutants or pathologies

Contexte : L’amiante est maintenant une cause avérée de cancer du larynx, mais le rôle des expositions professionnelles dans la survenue des cancers des voies aéro-digestives supérieures (VADS) reste largement méconnu. Bien que plusieurs études aient rapporté des associations entre les expositions professionnelles et ces cancers, il est difficile de synthétiser les résultats et d’en tirer des conclusions définitives. Objectifs : (1) Le premier objectif est de synthétiser les données épidémiologiques disponibles sur les associations entre les cancers de la cavité buccale, du pharynx et du larynx (les cancers des VADS les plus fréquents) et les expositions professionnelles ; (2) le second objectif est, à partir des données d’une large étude cas-témoins, l'étude Icare, d'identifier les professions ou industries présentant des risques élevés de cancer des VADS, puis d'étudier le rôle de certaines expositions professionnelles suspectées (amiante, laines minérales (LM), poussières de ciment, silice). Matériel et méthodes : (1) Une recherche bibliographique et des séries de méta-analyses ont été réalisées pour certaines expositions professionnelles suspectées. (2) L'étude Icare est une étude cas-témoins en population générale française incluant 2415 cas de cancer des VADS et 3555 témoins. L'historique professionnel complet des sujets a été recueilli, avec une description détaillée de chaque emploi exercé. L'évaluation des expositions à l'amiante, aux LM, aux poussières de ciment et à la silice sur l’ensemble de la vie professionnelle a été réalisée à l'aide des matrices emplois-expositions développées dans le cadre du programme Matgéné (Institut de Veille Sanitaire). Résultats : Des méta-risques relatifs (méta-RR) significativement augmentés ont été observés pour le cancer du larynx et les expositions aux hydrocarbures aromatiques polycycliques (HAP), aux gaz d'échappement de moteurs, au travail dans l'industrie textile et dans l'industrie du caoutchouc, et pour le cancer de la cavité buccale et du pharynx et les expositions à l'amiante, aux HAP et aux gaz d'échappement de moteurs. Des augmentations de risque significatives ont été observées pour plusieurs professions et secteurs d'activité exposant notamment aux nuisances présentant des méta-RR augmentés. Les résultats des analyses par nuisance confirment l'association entre l'amiante et le cancer du larynx et suggèrent également une association avec les cancers de la cavité buccale et du pharynx. Une association avec les poussières de ciment est également suggérée. Les résultats ne sont pas en faveur d’un rôle de l’exposition aux LM et à la silice.Conclusion : Ce travail renforce l’hypothèse d’un rôle des expositions professionnelles dans la survenue des cancers des VADS. Dans l’ensemble, nos résultats suggèrent des associations entre le risque de cancer des VADS et les expositions à l'amiante, aux HAP, aux poussières de ciment, et le travail dans l'industrie du caoutchouc. / Background: There is sufficient evidence that asbestos causes laryngeal cancer, but overall the role of occupational exposures in the etiology of head and neck cancer (HNC) remains largely unknown. Although several studies have reported associations between occupational exposures and HNC, it is difficult to draw firm conclusions. Objectives: (1) To summarize available epidemiologic data on occupational exposures and cancers of the oral cavity (OC), pharynx and larynx (the most frequent HNC); (2) using data from a large case-control study, to identify occupations and industries with an increased risk of HNC, then to investigate the role of some suspected occupational exposures (asbestos, mineral wools (MW), cement dust, silica). Methods: (1) A literature research and a series of meta-analyses were performed. (2) The Icare study is a French population-based case-control study including 2415 HNC cases and 3555 controls. Complete and detailed occupational histories were collected. Analyses by job title were conducted. Job exposure matrices, developed by the Occupational Health Department of the French Institute for Public Health Surveillance (InVS), were used to assess lifetime occupational exposure to asbestos, MW, cement dust and silica. Results: Significantly increased meta-relative risks (meta-RR) were obtained considering laryngeal cancer and exposures to polycyclic aromatic hydrocarbons (PAH), engine exhausts, working in the textile industry and the rubber industry, and for cancer of the OC and pharynx and exposures to asbestos, PAH and engine exhausts. Significantly increased risks were found for several jobs and industries, some of them entailing exposures to agents for which meta-RR were increased. Analyses for specific occupational exposures confirmed the association between asbestos and laryngeal cancer and showed an association with the risk of oral cavity and pharyngeal cancer. An association with exposure to cement dust was also suggested. The results did not support an association between HNC risk and exposure to MW or silica. Conclusion: This work emphasizes the role of occupational exposures in HNC. Overall, our results suggest associations between HNC and exposure to asbestos, PAH, cement dust, and work in the rubber industry.