O mercúrio (Hg) é um metal tóxico extensamente estudado em todo mundo, distribuído no ambiente a partir de fontes naturais ou antropogênicas e que oferece risco a população por ser altamente biocumulativo e possuir efeitos nocivos à saúde. Até algum tempo atrás, acreditava-se que a principal fonte de exposição ao Hg na Região Amazônica, decorria do uso deste metal para amalgamação de ouro nos garimpos da região. No entanto, o Hg é encontrado naturalmente nos solos da Região Amazônica e ao atingir os sistemas aquáticos, favorecidos principalmente pela erosão e pelas chuvas, passa por um processo de metilação catalisada por microorganismos, dando origem à forma orgânica do metal, o metilmercúrio (MeHg). Esta forma do metal se acumula no sedimento dos rios e em peixes representando atualmente a principal fonte de exposição ao mercúrio em população ribeirinha. Os biomarcadores de exposição ao Hg são freqüentemente utilizados para identificar e estimar o risco em que um indivíduo ou uma população está exposta. No entanto, pouco se conhece a respeito das variações inter-individuais de cada um deles. Neste sentido, este estudo teve como objetivo avaliar as variações inter-individuais em biomarcadores de exposição ao Hg em uma população ribeirinha do rio Tapajós, Pará. Para tal, 410 ribeirinhos, residentes em 12 comunidades ao longo do rio Tapajós, no estado do Pará participaram do estudo. Foram determinadas as concentrações de mercúrio total (THg) em sangue total, plasma, eritrócito, urina e de IHg e MeHg em cabelo dos voluntários. As concentrações de THg no sangue total variaram de 1,7 a 288,9 µg/L e no plasma de 0,2 a 40,0 µg/L. A concentração de THg no plasma apresentou uma alta correlação com as concentrações de THg em sangue total (r=0,7529, p<0,0001). A fração plasmática de THg variou 0,5% a 61% e não apresentou qualquer correlação com as concentrações de THg em sangue total (r= 0,06284, p=0,2041), indicando que a mobilização de THg para o plasma não ocorre devido à saturação dos eritrócitos. A distribuição de THg entre eritrócitos e plasma observada nesta população, é diferente do que foi observado em outros estudos sobre exposição à MeHg. As concentrações de THg no cabelo variaram de 0,97 a 62,4 µg/g e apresentaram uma correlação muito forte com as concentraçoes no sangue (r=0,8718, p<0,0001) indicando que as concentrações de THg no cabelo refletem as concentrações de THg no sangue. Também foi observada uma forte correlação entre as concentrações de MeHg e de IHg no cabelo (r=0,8979, p<0,0001), confirmando as informações da literatura, que sugerem que a fração de IHg no cabelo se deve a demetilação no sangue, no folículo capilar ou no preparo da amostra e análise. Em relação a razão entre concentração de mercúrio entre cabelo e sangue, observamos uma elevada variação entre os indivíduos, de 1:13 a 1:13274. Entre as mulheres observamos que esta variação ocorre de acordo com a idade. Resultados preliminares apontam para uma considerável variação inter-individual nos biomarcadores de exposição na população em estudo, indicando a necessidade de se identificar os fatores que influenciam este achado. Considerando a cinética do mercúrio, podemos concluir que estas variações inter-individuais na fração plasmática, podem alterar a taxa de eliminação do Hg e também os efeitos tóxicos decorrentes da exposição. Palavras- Chave: biomarcadores de exposição, mercúrio, variações inter-individuais / Mercury (Hg) is a toxic metal widely studied worldwide. In the atmosphere Hg may occur due to natural or anthropogenic sources. It offers risk to the population due to be highly bioaccumulated and to cause harmful effects to humans. Gold Mining activities were considered in the past the main sources of Hg contamination in the Amazon region. However, new findings indicated that Hg is naturally found in the soils of the Amazon area. When reaching the aquatic systems, facilitated mainly by the erosion and for the rains, the inorganic mercury is methylated, by microorganisms, forming the more toxic form methylmercury (MeHg). This form of the metal accumulates in the sediment of the rivers and in fish, meaning the main exposure source of mercury to riverine population. Biomarkers of exposure to Hg (levels of Hg in blood, plasma, urine, hair) are frequently used to identify and to esteem the risk of an individual or a population to harmful effects. However, little it is known regarding the inter-individual variations of these biomarkers. In this sense, this study evaluated inter-individual variations in the biomarkers of exposure to mercury (Hg in plasma, blood, urine and hair) in a riverside population (Tapajós river, Pará). Volunteers (n=410), residents in 12 communities along the Tapajós river, in the state of Pará participated in the study. Total mercury (THg) levels were determinated in whole blood, plasma, red blood cells and urine and of IHg and MeHg in hair. The concentration of mercury ranged from 1.7 to 288.9 µg/L and of plasma from 0.2 to 40.0 µg/L. The concentration of THg in the plasma presented a high correlation with concentrations of Hg in total blood (r=0.7529, p<0.0001). The plasmatic fraction of THg ranged from 0.5% to 61% and did not present any correlation with the concentrations of THg in the whole blood (r= 0.06284 p=0.2041), indicating that the mobilization of Hg to the plasma does not occur to the saturation of the red blood cells. The distribution of THg between red blood cells and plasma observed in this population is in disagreement when compared to other populations. The mercury concentrations in hair ranged from 0.97 to 62.4 µg/g and presented a very strong correlation with the whole blood (r=0.8718, p<0.0001), indicating that the concentrations of Hg in hair reflect the concentrations of THg in blood. Also a strong correlation was observed among the concentrations of MeHg and of IHg in hair (r=0.8979, p<0.0001), confirming the information in the literature, that suggest the fraction of IHg in hair is due to demethylation process or by sample preparation and analysis. Regarding the ratio between concentration of mercury between hair and blood, we observed a high variation between individuals, ranged from 1:13 to 1:13274. Among the women we observed this variation occurring according to age. In conclusion our results together demonstrated a considerable inter-individual variation in the biomarkers of exposure to mercury in the study population, demonstrating probably a different rate of biotransformation and elimination of Hg in this population. Then, future studies are necessary to elucidate factors are influencing this variation.

O mercúrio (Hg) é um metal tóxico extensamente estudado em todo mundo, distribuído no ambiente a partir de fontes naturais ou antropogênicas e que oferece risco a população por ser altamente biocumulativo e possuir efeitos nocivos à saúde. Até algum tempo atrás, acreditava-se que a principal fonte de exposição ao Hg na Região Amazônica, decorria do uso deste metal para amalgamação de ouro nos garimpos da região. No entanto, o Hg é encontrado naturalmente nos solos da Região Amazônica e ao atingir os sistemas aquáticos, favorecidos principalmente pela erosão e pelas chuvas, passa por um processo de metilação catalisada por microorganismos, dando origem à forma orgânica do metal, o metilmercúrio (MeHg). Esta forma do metal se acumula no sedimento dos rios e em peixes representando atualmente a principal fonte de exposição ao mercúrio em população ribeirinha. Os biomarcadores de exposição ao Hg são freqüentemente utilizados para identificar e estimar o risco em que um indivíduo ou uma população está exposta. No entanto, pouco se conhece a respeito das variações inter-individuais de cada um deles. Neste sentido, este estudo teve como objetivo avaliar as variações inter-individuais em biomarcadores de exposição ao Hg em uma população ribeirinha do rio Tapajós, Pará. Para tal, 410 ribeirinhos, residentes em 12 comunidades ao longo do rio Tapajós, no estado do Pará participaram do estudo. Foram determinadas as concentrações de mercúrio total (THg) em sangue total, plasma, eritrócito, urina e de IHg e MeHg em cabelo dos voluntários. As concentrações de THg no sangue total variaram de 1,7 a 288,9 µg/L e no plasma de 0,2 a 40,0 µg/L. A concentração de THg no plasma apresentou uma alta correlação com as concentrações de THg em sangue total (r=0,7529, p<0,0001). A fração plasmática de THg variou 0,5% a 61% e não apresentou qualquer correlação com as concentrações de THg em sangue total (r= 0,06284, p=0,2041), indicando que a mobilização de THg para o plasma não ocorre devido à saturação dos eritrócitos. A distribuição de THg entre eritrócitos e plasma observada nesta população, é diferente do que foi observado em outros estudos sobre exposição à MeHg. As concentrações de THg no cabelo variaram de 0,97 a 62,4 µg/g e apresentaram uma correlação muito forte com as concentraçoes no sangue (r=0,8718, p<0,0001) indicando que as concentrações de THg no cabelo refletem as concentrações de THg no sangue. Também foi observada uma forte correlação entre as concentrações de MeHg e de IHg no cabelo (r=0,8979, p<0,0001), confirmando as informações da literatura, que sugerem que a fração de IHg no cabelo se deve a demetilação no sangue, no folículo capilar ou no preparo da amostra e análise. Em relação a razão entre concentração de mercúrio entre cabelo e sangue, observamos uma elevada variação entre os indivíduos, de 1:13 a 1:13274. Entre as mulheres observamos que esta variação ocorre de acordo com a idade. Resultados preliminares apontam para uma considerável variação inter-individual nos biomarcadores de exposição na população em estudo, indicando a necessidade de se identificar os fatores que influenciam este achado. Considerando a cinética do mercúrio, podemos concluir que estas variações inter-individuais na fração plasmática, podem alterar a taxa de eliminação do Hg e também os efeitos tóxicos decorrentes da exposição. Palavras- Chave: biomarcadores de exposição, mercúrio, variações inter-individuais / Mercury (Hg) is a toxic metal widely studied worldwide. In the atmosphere Hg may occur due to natural or anthropogenic sources. It offers risk to the population due to be highly bioaccumulated and to cause harmful effects to humans. Gold Mining activities were considered in the past the main sources of Hg contamination in the Amazon region. However, new findings indicated that Hg is naturally found in the soils of the Amazon area. When reaching the aquatic systems, facilitated mainly by the erosion and for the rains, the inorganic mercury is methylated, by microorganisms, forming the more toxic form methylmercury (MeHg). This form of the metal accumulates in the sediment of the rivers and in fish, meaning the main exposure source of mercury to riverine population. Biomarkers of exposure to Hg (levels of Hg in blood, plasma, urine, hair) are frequently used to identify and to esteem the risk of an individual or a population to harmful effects. However, little it is known regarding the inter-individual variations of these biomarkers. In this sense, this study evaluated inter-individual variations in the biomarkers of exposure to mercury (Hg in plasma, blood, urine and hair) in a riverside population (Tapajós river, Pará). Volunteers (n=410), residents in 12 communities along the Tapajós river, in the state of Pará participated in the study. Total mercury (THg) levels were determinated in whole blood, plasma, red blood cells and urine and of IHg and MeHg in hair. The concentration of mercury ranged from 1.7 to 288.9 µg/L and of plasma from 0.2 to 40.0 µg/L. The concentration of THg in the plasma presented a high correlation with concentrations of Hg in total blood (r=0.7529, p<0.0001). The plasmatic fraction of THg ranged from 0.5% to 61% and did not present any correlation with the concentrations of THg in the whole blood (r= 0.06284 p=0.2041), indicating that the mobilization of Hg to the plasma does not occur to the saturation of the red blood cells. The distribution of THg between red blood cells and plasma observed in this population is in disagreement when compared to other populations. The mercury concentrations in hair ranged from 0.97 to 62.4 µg/g and presented a very strong correlation with the whole blood (r=0.8718, p<0.0001), indicating that the concentrations of Hg in hair reflect the concentrations of THg in blood. Also a strong correlation was observed among the concentrations of MeHg and of IHg in hair (r=0.8979, p<0.0001), confirming the information in the literature, that suggest the fraction of IHg in hair is due to demethylation process or by sample preparation and analysis. Regarding the ratio between concentration of mercury between hair and blood, we observed a high variation between individuals, ranged from 1:13 to 1:13274. Among the women we observed this variation occurring according to age. In conclusion our results together demonstrated a considerable inter-individual variation in the biomarkers of exposure to mercury in the study population, demonstrating probably a different rate of biotransformation and elimination of Hg in this population. Then, future studies are necessary to elucidate factors are influencing this variation.

Contexte - La variation interindividuelle de la réponse aux corticostéroïdes (CS) est un problème important chez les patients atteints de maladies inflammatoires d’intestin. Ce problème est bien plus accentué chez les enfants avec la prévalence de la corticodépendance extrêmement (~40 %) élevée. La maladie réfractaire au CS a des répercussions sur le développement et le bien-être physique et psychologique des patients et impose des coûts médicaux élevés, particulièrement avec la maladie active comparativement à la maladie en rémission, le coût étant 2-3 fois plus élevé en ambulatoire et 20 fois plus élevé en hôpital. Il est ainsi primordial de déterminer les marqueurs prédictifs de la réponse aux CS. Les efforts précédents de découvrir les marqueurs cliniques et démographiques ont été équivoques, ce qui souligne davantage le besoin de marqueurs moléculaires. L'action des CS se base sur des processus complexes déterminés génétiquement. Deux gènes, le ABCB1, appartenant à la famille des transporteurs transmembraneaux, et le NR3C1, encodant le récepteur glucocorticoïde, sont des éléments importants des voies métaboliques. Nous avons postulé que les variations dans ces gènes ont un rôle dans la variabilité observée de la réponse aux CS et pourraient servir en tant que les marqueurs prédictifs. Objectifs - Nous avons visé à: (1) examiner le fardeau de la maladie réfractaire aux CS chez les enfants avec la maladie de Crohn (MC) et le rôle des caractéristiques cliniques et démographiques potentiellement liés à la réponse; (2) étudier l'association entre les variantes d'ADN de gène ABCB1 et la réponse aux CS; (3) étudier les associations entre les variantes d'ADN de gène NR3C1 et la réponse aux CS. Méthodes - Afin d’atteindre ces objectifs, nous avons mené une étude de cohorte des patients recrutés dans deux cliniques pédiatriques tertiaires de gastroentérologie à l’Ottawa (CHEO) et à Montréal (HSJ). Les patients avec la MC ont été diagnostiqués avant l'âge de 18 ans selon les critères standard radiologiques, endoscopiques et histopathologiques. La corticorésistance et la corticodépendance ont été définies en adaptant les critères reconnus. L’ADN, acquise soit du sang ou de la salive, était génotypée pour des variations à travers de gènes ABCB1 et NR3C1 sélectionnées à l’aide de la méthodologie de tag-SNP. La fréquence de la corticorésistance et la corticodépendance a été estimée assumant une distribution binomiale. Les associations entre les variables cliniques/démographiques et la réponse aux CS ont été examinées en utilisant la régression logistique en ajustant pour des variables potentielles de confusion. Les associations entre variantes génétiques de ABCB1 et NR3C1 et la réponse aux CS ont été examinées en utilisant la régression logistique assumant différents modèles de la transmission. Les associations multimarqueurs ont été examinées en utilisant l'analyse de haplotypes. Les variantes nongénotypées ont été imputées en utilisant les données de HAPMAP et les associations avec SNPs imputés ont été examinées en utilisant des méthodes standard. Résultats - Parmi 645 patients avec la MC, 364 (56.2%) ont reçu CS. La majorité de patients étaient des hommes (54.9 %); présentaient la maladie de l’iléocôlon (51.7%) ou la maladie inflammatoire (84.6%) au diagnostic et étaient les Caucasiens (95.6 %). Huit pourcents de patients étaient corticorésistants et 40.9% - corticodépendants. Le plus bas âge au diagnostic (OR=1.34, 95% CI: 1.03-3.01, p=0.040), la maladie cœxistante de la région digestive supérieure (OR=1.35, 95% CI: 95% CI: 1.06-3.07, p=0.031) et l’usage simultané des immunomodulateurs (OR=0.35, 95% CI: 0.16-0.75, p=0.007) ont été associés avec la corticodépendance. Un total de 27 marqueurs génotypés à travers de ABCB1 (n=14) et NR3C1 (n=13) ont été en l'Équilibre de Hardy-Weinberg, à l’exception d’un dans le gène NR3C1 (rs258751, exclu). Dans ABCB1, l'allèle rare de rs2032583 (OR=0.56, 95% CI: 0.34-0.95, p=0.029) et génotype hétérozygote (OR=0.52, 95% CI: 0.28-0.95 p=0.035) ont été négativement associes avec la dépendance de CS. Un haplotype à 3 marqueurs, comprenant le SNP fonctionnel rs1045642 a été associé avec la dépendance de CS (p empirique=0.004). 24 SNPs imputés introniques et six haplotypes ont été significativement associés avec la dépendance de CS. Aucune de ces associations n'a cependant maintenu la signification après des corrections pour des comparaisons multiples. Dans NR3C1, trois SNPs: rs10482682 (OR=1.43, 95% CI: 0.99-2.08, p=0.047), rs6196 (OR=0.55, 95% CI: 0.31-0.95, p=0.024), et rs2963155 (OR=0.64, 95% CI: 0.42-0.98, p=0.039), ont été associés sous un modèle additif, tandis que rs4912911 (OR=0.37, 95% CI: 0.13-1.00, p=0.03) et rs2963156 (OR=0.32, 95% CI: 0.07-1.12, p=0.047) - sous un modèle récessif. Deux haplotypes incluant ces 5 SNPs (AAACA et GGGCG) ont été significativement (p=0.006 et 0.01 empiriques) associés avec la corticodépendance. 19 SNPs imputés ont été associés avec la dépendance de CS. Deux haplotypes multimarqueurs (p=0.001), incluant les SNPs génotypés et imputés, ont été associés avec la dépendance de CS. Conclusion - Nos études suggèrent que le fardeau de la corticodépendance est élevé parmi les enfants avec le CD. Les enfants plus jeunes au diagnostic et ceux avec la maladie coexistante de la région supérieure ainsi que ceux avec des variations dans les gènes ABCB1 et NR3C1 étaient plus susceptibles de devenir corticodépendants. / Background - Inter-individual variation in response to treatment by corticosteroids (CS) is an important problem in the management of inflammatory bowel disease (IBD) patient’s. This problem is even more prominent in children, the prevalence of steroid dependence (~40%) in whom is extremely high. Steroid refractoriness has a considerable impact on the physical and psychological development of these children, also imposing high medical costs related to treatment. Active disease, as opposed to quiescent, increases medical costs 2-3 times in ambulatory patients and 20 times in hospitalized cases. Identifying markers that could predict steroid response is therefore a high clinical priority. Previous attempts to investigate potential clinical and demographic markers have been equivocal, highlighting the need for further investigations of other predictive markers. It is well known that the action of CS entails complex processes controlled by genetic factors. Two genes, the ABCB1 gene, which belongs to the family of trans-membrane transporters, and the NR3C1 gene, coding for the glucocorticoid receptor, are major elements of the pathway. We postulated that inter-individual variations in these genes may play a role in the observed variability of the response to CS and could serve as potential predictors. Objectives - We aimed to: (1) examine the burden of steroid refractoriness in children diagnosed with CD and explore the potential clinical/demographic factors related to CS response; (2) study the association between DNA variants in the ABCB1 gene and CS response; (3) investigate the associations between DNA variants in the NR3C1 gene and CS response. Methods - We investigated these objectives in a cohort of CD patients recruited from two tertiary paediatric gastroenterology clinics from Ottawa (CHEO) and Montreal (HSJ). CD patients diagnosed prior to age 18 using standard clinical, radiological, endoscopic and histopathological criteria were included. Published criteria were adapted to define CS-resistance and dependence. DNA acquired from blood and/or saliva was genotyped for variations across the ABCB1 and NR3C1 genes selected using the tag-SNP methodology. The frequencies of steroid resistance and dependence were estimated assuming a binomial distribution. Associations between clinical/demographic variables and steroid responses were examined using logistic regression modeling after accounting for potential confounding variables. Associations between ABCB1 and NR3C1 genes’ variants and steroid responses were examined using logistic regression assuming different models of inheritance. Multi-marker associations were examined via haplotype analysis. Un-genotyped variants in the genes were imputed using HAPMAP data as the reference panel and associations with imputed SNPs examined using standard methods. Results - Among 645 CD patients diagnosed at the study centers, 364 (56.2%) received corticosteroids during the first year since diagnosis. The majority of patients were male (54.9%), had inflammatory (84.6%), ileo-colonic (51.7%) disease phenotypes at diagnosis and were Caucasians (95.6%). Eight percent of patients developed CS-resistance and 40.9% became CS-dependent. Younger age at diagnosis (OR=1.34, 95% CI: 1.03-3.01, p=0.040), coexisting upper digestive tract involvement (OR=1.35, 95% CI: 1.06-3.07, p=0.031) and concomitant immunomodulators use (OR=0.35, 95% CI: 0.16-0.75, p=0.007) were significantly associated with CS-dependency in multivariate analysis. From among the 27 markers genotyped across the ABCB1 (n=14) and NR3C1 genes (n=13), all except one in NR3C1 gene (rs258751, excluded) were in Hardy-Weinberg Equilibrium. For ABCB1, the rare allele of rs2032583 (OR=0.56, 95% CI: 0.34-0.95, p=0.029) and heterozygous genotype (OR=0.52, 95% CI: 0.28-0.95, p=0.035) conferred protection from CS dependency. A 3-marker haplotype including the functional SNP rs1045642 was associated with CS-dependence (empiric p-value=0.004). On imputation 24 intronic SNPs and six haplotypes were statistically significantly associated with CS dependence. None of these associations however maintained significance after corrections for multiple comparisons. For the NR3C1 gene 3 SNPs, rs10482682 (OR=1.43, 95% CI: 0.99-2.08, p=0.047), rs6196 (OR=0.55, 95% CI: 0.31-0.95, p=0.024), and rs2963155 (OR=0.64, 95% CI: 0.42-0.98, p=0.039), showed associations under an additive model whereas rs4912911 (OR=0.37, 95% CI: 0.13-1.00, p=0.03) and rs2963156 (OR=0.32, 95% CI: 0.07-1.12, p=0.047) showed associations under a recessive model. Two haplotypes encompassing these 5 SNPs (AAACA and GGGCG) were significantly (empirical p=0.006 and 0.01 respectively) were associated with CS-dependence. On imputation 19 SNPs were associated with CS-dependence. Two multi-marker haplotypes (p-values=0.001 each) including genotyped and imputed SNPs conferred susceptibility for CS-dependency. Conclusions - Our studies suggest that the burden of steroid dependence is high among children with CD. Children diagnosed at a younger age, those with co-existent upper tract disease and with variations in the ABCB1 and NR3C1 genes were more likely to become CS dependent.

Contexte - La variation interindividuelle de la réponse aux corticostéroïdes (CS) est un problème important chez les patients atteints de maladies inflammatoires d’intestin. Ce problème est bien plus accentué chez les enfants avec la prévalence de la corticodépendance extrêmement (~40 %) élevée. La maladie réfractaire au CS a des répercussions sur le développement et le bien-être physique et psychologique des patients et impose des coûts médicaux élevés, particulièrement avec la maladie active comparativement à la maladie en rémission, le coût étant 2-3 fois plus élevé en ambulatoire et 20 fois plus élevé en hôpital. Il est ainsi primordial de déterminer les marqueurs prédictifs de la réponse aux CS. Les efforts précédents de découvrir les marqueurs cliniques et démographiques ont été équivoques, ce qui souligne davantage le besoin de marqueurs moléculaires. L'action des CS se base sur des processus complexes déterminés génétiquement. Deux gènes, le ABCB1, appartenant à la famille des transporteurs transmembraneaux, et le NR3C1, encodant le récepteur glucocorticoïde, sont des éléments importants des voies métaboliques. Nous avons postulé que les variations dans ces gènes ont un rôle dans la variabilité observée de la réponse aux CS et pourraient servir en tant que les marqueurs prédictifs. Objectifs - Nous avons visé à: (1) examiner le fardeau de la maladie réfractaire aux CS chez les enfants avec la maladie de Crohn (MC) et le rôle des caractéristiques cliniques et démographiques potentiellement liés à la réponse; (2) étudier l'association entre les variantes d'ADN de gène ABCB1 et la réponse aux CS; (3) étudier les associations entre les variantes d'ADN de gène NR3C1 et la réponse aux CS. Méthodes - Afin d’atteindre ces objectifs, nous avons mené une étude de cohorte des patients recrutés dans deux cliniques pédiatriques tertiaires de gastroentérologie à l’Ottawa (CHEO) et à Montréal (HSJ). Les patients avec la MC ont été diagnostiqués avant l'âge de 18 ans selon les critères standard radiologiques, endoscopiques et histopathologiques. La corticorésistance et la corticodépendance ont été définies en adaptant les critères reconnus. L’ADN, acquise soit du sang ou de la salive, était génotypée pour des variations à travers de gènes ABCB1 et NR3C1 sélectionnées à l’aide de la méthodologie de tag-SNP. La fréquence de la corticorésistance et la corticodépendance a été estimée assumant une distribution binomiale. Les associations entre les variables cliniques/démographiques et la réponse aux CS ont été examinées en utilisant la régression logistique en ajustant pour des variables potentielles de confusion. Les associations entre variantes génétiques de ABCB1 et NR3C1 et la réponse aux CS ont été examinées en utilisant la régression logistique assumant différents modèles de la transmission. Les associations multimarqueurs ont été examinées en utilisant l'analyse de haplotypes. Les variantes nongénotypées ont été imputées en utilisant les données de HAPMAP et les associations avec SNPs imputés ont été examinées en utilisant des méthodes standard. Résultats - Parmi 645 patients avec la MC, 364 (56.2%) ont reçu CS. La majorité de patients étaient des hommes (54.9 %); présentaient la maladie de l’iléocôlon (51.7%) ou la maladie inflammatoire (84.6%) au diagnostic et étaient les Caucasiens (95.6 %). Huit pourcents de patients étaient corticorésistants et 40.9% - corticodépendants. Le plus bas âge au diagnostic (OR=1.34, 95% CI: 1.03-3.01, p=0.040), la maladie cœxistante de la région digestive supérieure (OR=1.35, 95% CI: 95% CI: 1.06-3.07, p=0.031) et l’usage simultané des immunomodulateurs (OR=0.35, 95% CI: 0.16-0.75, p=0.007) ont été associés avec la corticodépendance. Un total de 27 marqueurs génotypés à travers de ABCB1 (n=14) et NR3C1 (n=13) ont été en l'Équilibre de Hardy-Weinberg, à l’exception d’un dans le gène NR3C1 (rs258751, exclu). Dans ABCB1, l'allèle rare de rs2032583 (OR=0.56, 95% CI: 0.34-0.95, p=0.029) et génotype hétérozygote (OR=0.52, 95% CI: 0.28-0.95 p=0.035) ont été négativement associes avec la dépendance de CS. Un haplotype à 3 marqueurs, comprenant le SNP fonctionnel rs1045642 a été associé avec la dépendance de CS (p empirique=0.004). 24 SNPs imputés introniques et six haplotypes ont été significativement associés avec la dépendance de CS. Aucune de ces associations n'a cependant maintenu la signification après des corrections pour des comparaisons multiples. Dans NR3C1, trois SNPs: rs10482682 (OR=1.43, 95% CI: 0.99-2.08, p=0.047), rs6196 (OR=0.55, 95% CI: 0.31-0.95, p=0.024), et rs2963155 (OR=0.64, 95% CI: 0.42-0.98, p=0.039), ont été associés sous un modèle additif, tandis que rs4912911 (OR=0.37, 95% CI: 0.13-1.00, p=0.03) et rs2963156 (OR=0.32, 95% CI: 0.07-1.12, p=0.047) - sous un modèle récessif. Deux haplotypes incluant ces 5 SNPs (AAACA et GGGCG) ont été significativement (p=0.006 et 0.01 empiriques) associés avec la corticodépendance. 19 SNPs imputés ont été associés avec la dépendance de CS. Deux haplotypes multimarqueurs (p=0.001), incluant les SNPs génotypés et imputés, ont été associés avec la dépendance de CS. Conclusion - Nos études suggèrent que le fardeau de la corticodépendance est élevé parmi les enfants avec le CD. Les enfants plus jeunes au diagnostic et ceux avec la maladie coexistante de la région supérieure ainsi que ceux avec des variations dans les gènes ABCB1 et NR3C1 étaient plus susceptibles de devenir corticodépendants. / Background - Inter-individual variation in response to treatment by corticosteroids (CS) is an important problem in the management of inflammatory bowel disease (IBD) patient’s. This problem is even more prominent in children, the prevalence of steroid dependence (~40%) in whom is extremely high. Steroid refractoriness has a considerable impact on the physical and psychological development of these children, also imposing high medical costs related to treatment. Active disease, as opposed to quiescent, increases medical costs 2-3 times in ambulatory patients and 20 times in hospitalized cases. Identifying markers that could predict steroid response is therefore a high clinical priority. Previous attempts to investigate potential clinical and demographic markers have been equivocal, highlighting the need for further investigations of other predictive markers. It is well known that the action of CS entails complex processes controlled by genetic factors. Two genes, the ABCB1 gene, which belongs to the family of trans-membrane transporters, and the NR3C1 gene, coding for the glucocorticoid receptor, are major elements of the pathway. We postulated that inter-individual variations in these genes may play a role in the observed variability of the response to CS and could serve as potential predictors. Objectives - We aimed to: (1) examine the burden of steroid refractoriness in children diagnosed with CD and explore the potential clinical/demographic factors related to CS response; (2) study the association between DNA variants in the ABCB1 gene and CS response; (3) investigate the associations between DNA variants in the NR3C1 gene and CS response. Methods - We investigated these objectives in a cohort of CD patients recruited from two tertiary paediatric gastroenterology clinics from Ottawa (CHEO) and Montreal (HSJ). CD patients diagnosed prior to age 18 using standard clinical, radiological, endoscopic and histopathological criteria were included. Published criteria were adapted to define CS-resistance and dependence. DNA acquired from blood and/or saliva was genotyped for variations across the ABCB1 and NR3C1 genes selected using the tag-SNP methodology. The frequencies of steroid resistance and dependence were estimated assuming a binomial distribution. Associations between clinical/demographic variables and steroid responses were examined using logistic regression modeling after accounting for potential confounding variables. Associations between ABCB1 and NR3C1 genes’ variants and steroid responses were examined using logistic regression assuming different models of inheritance. Multi-marker associations were examined via haplotype analysis. Un-genotyped variants in the genes were imputed using HAPMAP data as the reference panel and associations with imputed SNPs examined using standard methods. Results - Among 645 CD patients diagnosed at the study centers, 364 (56.2%) received corticosteroids during the first year since diagnosis. The majority of patients were male (54.9%), had inflammatory (84.6%), ileo-colonic (51.7%) disease phenotypes at diagnosis and were Caucasians (95.6%). Eight percent of patients developed CS-resistance and 40.9% became CS-dependent. Younger age at diagnosis (OR=1.34, 95% CI: 1.03-3.01, p=0.040), coexisting upper digestive tract involvement (OR=1.35, 95% CI: 1.06-3.07, p=0.031) and concomitant immunomodulators use (OR=0.35, 95% CI: 0.16-0.75, p=0.007) were significantly associated with CS-dependency in multivariate analysis. From among the 27 markers genotyped across the ABCB1 (n=14) and NR3C1 genes (n=13), all except one in NR3C1 gene (rs258751, excluded) were in Hardy-Weinberg Equilibrium. For ABCB1, the rare allele of rs2032583 (OR=0.56, 95% CI: 0.34-0.95, p=0.029) and heterozygous genotype (OR=0.52, 95% CI: 0.28-0.95, p=0.035) conferred protection from CS dependency. A 3-marker haplotype including the functional SNP rs1045642 was associated with CS-dependence (empiric p-value=0.004). On imputation 24 intronic SNPs and six haplotypes were statistically significantly associated with CS dependence. None of these associations however maintained significance after corrections for multiple comparisons. For the NR3C1 gene 3 SNPs, rs10482682 (OR=1.43, 95% CI: 0.99-2.08, p=0.047), rs6196 (OR=0.55, 95% CI: 0.31-0.95, p=0.024), and rs2963155 (OR=0.64, 95% CI: 0.42-0.98, p=0.039), showed associations under an additive model whereas rs4912911 (OR=0.37, 95% CI: 0.13-1.00, p=0.03) and rs2963156 (OR=0.32, 95% CI: 0.07-1.12, p=0.047) showed associations under a recessive model. Two haplotypes encompassing these 5 SNPs (AAACA and GGGCG) were significantly (empirical p=0.006 and 0.01 respectively) were associated with CS-dependence. On imputation 19 SNPs were associated with CS-dependence. Two multi-marker haplotypes (p-values=0.001 each) including genotyped and imputed SNPs conferred susceptibility for CS-dependency. Conclusions - Our studies suggest that the burden of steroid dependence is high among children with CD. Children diagnosed at a younger age, those with co-existent upper tract disease and with variations in the ABCB1 and NR3C1 genes were more likely to become CS dependent.