Recessão gengival é uma condição caracterizada pelo deslocamento apical da gengiva marginal expondo a superfície radicular, sendo associada a problemas estéticos e funcionais. O objetivo deste estudo foi determinar a prevalência, extensão, gravidade e indicadores de risco para recessão gengival em adultos acima de 35 anos de idade residentes na cidade de Porto Alegre. Este é um estudo observacional transversal de base populacional. Uma amostra representativa de 1023 indivíduos foi selecionada utilizando-se uma amostragem aleatória proporcional de múltiplos-estágios. Os indivíduos responderam a um questionário estruturado e receberam um exame clínico em quatro sítios por dente de todos os dentes presentes. Comandos analíticos que consideram a amostra complexa foram utilizados para gerar estimativas e em modelos de risco multivariados. Recessão ≥1 mm foi um achado universal (99,7% dos indivíduos). O percentual de indivíduos com pelo menos um dente com recessão ≥3 mm, ≥5 mm e ≥7 mm foi 75,4%, 40,7% e 12,5%, respectivamente. Quanto a sua extensão, 67,6%, 27,8%, 9,5% e 2,1% dos dentes por indivíduo mostraram recessão gengival ≥1 mm, ≥3 mm, ≥5 mm e ≥7 mm, respectivamente. Os dentes mais afetados foram incisivos centrais inferiores, segundos pré-molares inferiores e primeiros molares superiores. O avanço da idade, gênero masculino, exposição ao fumo, nível elevado de educação e pior higiene bucal foram indicadores de risco para recessão gengival. Variáveis relativas à higiene bucal não estiveram associadas com recessão gengival vestibular. Pode-se concluir que existe elevada prevalência de recessão gengival na população estudada, estando associada a diferentes fatores comportamentais e sociodemográficos. Diante disto, políticas públicas de saúde são necessárias para agir sobre os fatores etiológicos desta condição e suas consequências. / Gingival recession is a condition characterized by apical displacement of the gingival margin exposing the root surface, associated with aesthetic and functional problems. The aim of this study was to determine the prevalence, extent, severity and risk indicators of gingival recession in adults over 35 years of age living in the city of Porto Alegre. This is an observational cross-sectional population-based study. A representative sample of 1023 individuals was selected using a multistage stratified random sampling. The subjects answered a structured questionnaire and received a clinical examination at four sites per tooth of all teeth present. Analytical commands that consider the complex sample were used to generate estimates and multivariate risk models. Recession ≥ 1 mm was a universal finding (99.7% of subjects). The percentage of subjects with at least one tooth with recession ≥ 3 mm, ≥ 5 mm and ≥ 7 mm was 75.4%, 40.7% and 12.5%, respectively. In regards to extent, 67.6%, 27.8%, 9.5% and 2.1% of teeth per individual showed gingival recession ≥ 1 mm, ≥ 3 mm, ≥ 5 mm and ≥ 7 mm, respectively. The most affected teeth were mandibular central incisors, mandibular second premolars and maxillary first molars. Older age, male gender, smoking exposure, high education and poor oral hygiene were risk indicators for gingival recession. Variables related to oral hygiene were not associated with buccal gingival recession. It can be concluded that there is a high prevalence of gingival recession in this population and it is associated with different sociodemographic and behavioral factors. Given this, public health policies are needed to act on the etiological factors of this condition and its consequences.

Introdução: As doenças cardiovasculares (DCV) são a principal causa de morte no mundo, sendo a doença arterial coronária (DAC) a mais comum das DCV, juntamente com acidente vascular cerebral. O cálcio das artérias coronárias é um marcador da DAC subclínica (assintomática) e é preditivo de eventos coronários futuros. Vários fatores de risco cardiovascular contribuem para o desenvolvimento da calcificação da artéria coronária (CAC). Adicionalmente, os fatores dietéticos podem influenciar no processo de aterosclerose e CAC. Objetivo: Avaliar a CAC e sua associação com fatores de risco cardiovascular e ingestão dietética em homens assintomáticos vivendo em comunidade. Métodos: Estudo transversal. A amostra consistiu de 150 homens assintomáticos com idades entre 50 e 70 anos (idade média 58,2 ± 5,3 anos), que foram submetidos à tomografia computadorizada multidetectores (TCMD). A aterosclerose subclínica foi avaliada pela CAC de acordo com o método de Agatston, sendo os escores de cálcio classificados como ≤10 (sem evidência e CAC mínima) e >10 (CAC moderada e aumentada). A ingestão dietética foi avaliada através do Registro de Consumo Alimentar (RCA). O modelo multivariado de Regressão de Modified Poisson foi utilizado para avaliar os fatores de risco cardiovascular independentemente associados com a CAC moderada/aumentada, sendo estimados os efeitos do consumo de diversos nutrientes na prevalência de CAC moderada/aumentada ajustado para ingestão calórica e fatores de risco para CAC, através da razão de prevalências e intervalo de 95% de confiança. Resultados: A presença de CAC (escore de cálcio >0) foi identificada em 59,3% dos participantes. Na análise multivariada, os fatores independentemente associados com a CAC moderada/aumentada foram a história familiar (HF) de DAC prematura (RP=1,39; IC95% 1,03-1,88, p=0,029) e a atividade física (AF) <150 minutos/semana (RP=1,40; IC95% 1,01-1,93; p=0,045). O consumo de alguns nutrientes, também, mostrou-se associado à CAC moderada/aumentada, como o menor consumo de carboidratos (p=0,021) e o maior consumo de lipídeos (p=0,006), após o ajuste do modelo para a quantidade de calorias. Com a inclusão no modelo dos fatores de risco cardiovascular e escolaridade, os nutrientes associados à prevalência da CAC moderada/aumentada foram: carboidratos percentual (RP=0,98; IC95% 0,96-0,99; p=0,040), lipídeos percentual (RP=1,04; IC95% 1,01-1,07; p=0,005) e ácidos graxos saturados (AGS) percentual (RP=1,08; IC95% 1,02-1,14; p=0,013). Conclusões: Nesta amostra de adultos e idosos assintomáticos vivendo em comunidade, fatores de risco cardiovascular como HF de DAC prematura e baixa intensidade de AF estiveram associados, de forma independente, com a calcificação coronária moderada a aumentada. Analisando os fatores dietéticos, uma maior ingestão de lipídeos totais e de gorduras saturadas mostraram-se associadas com escores mais elevados de CAC, enquanto que a ingestão maior de carboidratos, em detrimento aos lipídeos, associou-se com escores mais baixos de CAC. Nossos resultados sugerem que esses fatores de risco devem ser mais considerados na avaliação clínica do risco cardiovascular global do paciente. / Introduction: Cardiovascular diseases (CVD) are the main cause of death in the world, being the coronary artery disease (CAD) the most common CVD. The calcium of the coronary arteries is a marker for subclinical (asymptomatic) CAD, and is predictive of future coronary events. A number of cardiovascular risk factors account for coronary artery calcification (CAC). In addition, dietary factors may influence the process of atherosclerosis and CAC. Goal: To assess CAC and its association with cardiovascular risk factors and dietary intake in community-dwelling asymptomatic men. Method: Cross-sectional study. The sample included 150 asymptomatic men with age ranging between 50 and 70 years (mean age 58.2 ± 5.3 years) submitted to multidetector computed tomography (MDCT). Subclinical atherosclerosis was measured by CAC in accordance with the Agatston method, with the scores classified as ≤10 (no evidence of, or mild CAC) and >10 (moderate and severe CAC). Dietary intake was assessed according to the Food consumption Register (RCA) method. The multivariate Modified Poisson regression model was used to assess cardiovascular risk associated with moderate/severe CAC and the effects of the intake of different nutrients were estimated for the prevalence of moderate/severe CAC, adjusted for calorie intake and CAC risk factors by means of prevalence ratios and 95% confidence intervals. Results: CAC (calcium score >0) was present in 59.3% of the subjects. In the multivariate analysis, factors independently associated with moderate/severe CAC included family history (FH) of early CAD (PR=1.39; 95%CI 1.03-1.88, p=0.029) and physical activity (PA) <150 minutes/week (PR=1.40; 95%CI 1.01-1.93; p=0.045). The intake of some nutrients was also associated with moderate/severe CAC, such as lower carbohydrate intake (p=0.021) and higher lipid intake (p=0.006), after model adjustment for the amount of calories. Once the cardiovascular risk factors and schooling were included in the model, the nutrients associated with the prevalence of moderate/severe CAC were: percentage of carbohydrates (PR=0.98; 95%CI 0.96-0.99; p=0.040), percentage of lipids (PR=1.04; 95%CI 1.01-1.07; p=0.005), and percentage of saturated fatty acids (SFA) (PR=1.08; 95%CI 1.02-1.14; p=0.013). Conclusions: In the sample of community-dwelling asymptomatic adults and older persons, cardiovascular risk factors such as FH of early CAD, and low-intensity PA were independently associated with moderate to severe coronary calcification. Analysis of dietary factors showed that higher intake of total lipids and saturated fats were associated with higher CAC scores, whereas higher intake of carbohydrates over lipids was associated with lower CAC scores. Our results indicate that these risk factors should be considered in the cardiovascular assessment of the patient.

Objectifs : Les cancers de la thyroïde représentent environ 2% de l‘ensemble des cancers dans les pays développés mais sont les cancers les plus fréquents des glandes endocrines. L’incidence des cancers de la thyroïde se caractérise par de fortes variations géographiques dans le monde, les taux les plus élevés étant observés en Nouvelle-Calédonie, et par une augmentation régulière au cours des dernières décennies dans les pays occidentaux. Cette augmentation d’incidence est en grande partie attribuée à l’évolution des pratiques médicales mais pourrait aussi être en partie liée à une évolution des modes de vie et/ou à des expositions environnementales. En dehors de l’exposition aux radiations ionisantes dans l’enfance, les facteurs de risque de cancer de la thyroïde restent très mal connus. L’objectif de ce travail était d’étudier le rôle des antécédents familiaux de pathologie de la thyroïde, des caractéristiques de la vie hormonale et reproductive, des caractéristiques anthropométriques et des expositions environnementales et professionnelles dans l’étiologie des cancers de la thyroïde. Méthode : Les analyses reposent principalement sur une étude cas-témoins en population menée en France métropolitaine (étude CATHY). Nous avons également utilisé les données d’une étude cas-témoins menée en Nouvelle-Calédonie pour étudier le rôle des antécédents familiaux dans cette zone géographique de forte incidence. L’étude CATHY repose sur 621 cas diagnostiqués dans trois départements français (Calvados, Marne, Ardennes) entre 2002 et 2007 et sur 706 témoins, appariés par fréquence aux cas sur l’âge, le sexe et le département de résidence. L’étude cas-témoins en Nouvelle-Calédonie repose sur 332 cas diagnostiqués entre 1985 et 1999, et sur 412 témoins appariés par fréquence aux cas sur l’âge et le sexe. Résultats : Les résultats montrent une augmentation du risque de cancer de la thyroïde chez les individus ayant des apparentés de premier degré atteints de cancer de la thyroïde ou de goitre multinodulaire, en France métropolitaine comme en Nouvelle-Calédonie. Parmi les facteurs hormonaux et reproductifs, un âge tardif aux premières règles, une ménopause précoce et une parité élevée étaient associés à une augmentation du risque de cancer de la thyroïde. Une relation dose-effet inverse était observée entre la durée cumulée des cycles menstruels au cours de la vie et le risque de cancer de la thyroïde. La prise d’une contraception orale était associée à une diminution du risque de cancer de la thyroïde, avec une relation dose-effet inverse entre la durée de la contraception orale et le risque de cancer de la thyroïde. La taille, l’indice de masse corporelle et la surface corporelle étaient positivement associés au risque de cancer de la thyroïde. Une augmentation du risque de cancer de la thyroïde était observée dans certains secteurs d’activité professionnelle comme l’industrie du papier, l’industrie du caoutchouc et des plastiques, le travail du bois, la réparation automobile et la métallurgie. / Objectives: Thyroid cancer accounts for 2% of all cancers in industrialized countries but is the most common endocrine cancer. There is a marked worldwide geographic variation in thyroid cancer incidence, the highest incidence rates being observed in New Caledonia, and a constant increase of thyroid cancer incidence has been observed during the past several decades. Although the evolution in medical practice probably accounts for some of the increase, other factors such as environmental and/or lifestyle factors may also play a role. However, there are few established risk factors of thyroid cancer apart from ionizing radiation exposure in childhood. We aimed to investigate the etiologic role of a family history of thyroid cancer, of hormonal and reproductive factors, anthropometric factors and environmental and occupational exposures in thyroid cancer. Methods: Analyses mainly rely on a population-based case-control study conducted in metropolitan France (“CATHY study”). Data of a case-control study conducted in New-Caledonia were also used to investigate the role of a family history of thyroid cancer as a risk factor for thyroid cancer in this area where thyroid cancer incidence is very high. The CATHY study is based on 621 cases diagnosed from 2002 to 2007 in three administrative areas: the Calvados, the Marne and the Ardennes, and 705 controls matched to cases by age, sex and residence area. The New Caledonian case-control study is based on 332 cases diagnosed form 1985 to 1999 and 412 controls matched to cases by age and sex. Results: A family history of thyroid cancer or multinodular goiter in first degree relatives was associated with an increased risk of thyroid cancer in France as well as in New Caledonia. Among hormonal and reproductive factors, a later age at menarche, a younger age at menopause and a elevated number of full-term pregnancies were associated with an increased risk of thyroid cancer. An inverse relationship was observed between the cumulative number of menstrual cycles and risk of thyroid cancer. Oral contraceptive use was associated with a lower risk of thyroid cancer, and an inverse relationship was observed between oral contraceptive use duration and risk of thyroid cancer. Height, body mass index and body surface area were positively associated with thyroid cancer risk. Analyses by occupation showed a higher risk of thyroid cancer for workers in leather industry, in plastics and rubber industry, in papermaking, in wood manufacturing, in automotive repair and in metallurgy.

De nombreux facteurs de risque cardiovasculaire comme l’hypercholestérolémie, l’hypertension, et le diabète sont comptés parmi les facteurs de risque modifiables les plus importants pour le déclin cognitif et la démence. L’exposition à ces facteurs de risque au cours de la vie en particulier avant l’âge de 65 ans ainsi que leur agrégation contribuent de manière plus importante au déclin cognitif. Peu d’études se sont intéressées aux mesures composites de risque cardiovasculaire par rapport à la fonction cognitive chez les sujets de moins de 65 ans. L’objectif de cette thèse était d’étudier l’association entre les mesures composites de risque et le déclin cognitif au cours de la phase précocedu vieillissement. Les données de la cohorte Whitehall II dans laquelle les fonctions cognitives ont été mesurées à trois reprises ont été utilisées pour étudier l’association entre le syndrome métabolique et deux scores de risque de Framingham (de maladie cardiovasculaire globale et d’AVC), et la fonction cognitive et le déclin cognitif sur 10 ans. Les scores de risque cardiovasculaire de Framingham ont aussi été comparés avec un score de risque de démence. De toutes les mesures composites de risque étudiées, les scores de risque de Framingham montraient la plus forteassociation avec le déclin cognitif. Un risque cardiovasculaire plus élevé était associé à un déclin plus rapide dans de multiples tests cognitifs dont la fluence verbale, le vocabulaire et la cognition globale. Ces résultats suggèrent d’une part qu’un risque cardiovasculaire plus élevé contribue au déclin cognitif dès la phase précoce de vieillissement et d’autre part que l’estimation du risque cardiovasculaire et son effet sur la fonction cognitive nécessite une approche multifactorielle.L’identification et la réduction de facteurs de risque cardiovasculaire peuvent avoir un impact important sur la réduction du déclin cognitif et de la démence. / Several cardiovascular disease risk factors including, dyslipidemia, high blood pressure, and diabetes have been proposed as important modifiable risk factors for cognitive decline and dementia. These risk factors often co-occur and their aggregation is associated with increased risk of cardiovascular disease and dementia. However, studies of composite measures of cardiovascular disease risk in relation to cognitive outcomes in non-elderly populations are scarce. The aim of this thesis was to examine composite measures of risk in relation to cognition and longitudinal cognitive change amongmiddle-aged adults. Data from the Whitehall II study were used to study the associations between the metabolic syndrome, two Framingham risk scores; the Framingham stroke and general cardiovascular disease risk scores, and cognition, based on three cognitive assessments over 10 years. In addition, these two (cardio)vascular risk scores were compared with the CAIDE dementia risk score. Of all composite measures of risk examined, the two Framingham risk scores were the best predictors of 10-year cognitive decline. Higher cardiovascular risk was associated with faster 10-year decline inmultiple cognitive tests including verbal fluency, vocabulary and global cognition. These results suggest that multiple cardiovascular disease risk factors contribute to cognitive decline starting in midlife and that multi-risk factor models such as cardiovascular risk scores may be better suited to assessing risk of cognitive decline. Early identification and treatment of cardiovascular disease risk factors may offer the possibility of markedly delaying or preventing cognitive decline.

Les femmes porteuses d’une mutation du gène BRCA1 ou BRCA2 ont un risque de cancer du sein (CS) très élevé dont les estimations varient beaucoup d’une étude à l’autre. L’objectif principal de cette étude est de mieux estimer le risque de CS associé aux gènes BRCA1/2 en tenant compte de la variabilité des mutations et des facteurs « environnementaux/style de vie » et de leur éventuelle interaction. Nous avons analysé les données de la cohorte GENEPSO composée de femmes porteuses d’une mutation du gène BRCA1 ou BRCA2 à l’aide d’un modèle de Cox pondéré. L’analyse des facteurs de risque gynéco-obstétrique et de « style de vie » a permis de mettre en évidence une association entre le risque de CS et les radiations ionisantes, la consommation de tabac, l’indice de masse corporelle, l’âge aux premières règles, la parité, les interruptions de grossesse, la contraception orale, la ménopause et les traitements hormonaux substitutifs. Cette étude confirme l’existence d’une zone centrale à moindre risque de CS dans les gènes BRCA1/2 et de décrire une nouvelle région à haut risque située dans la région 3’ du gène BRCA2. Cette étude montre également une interaction entre la localisation des mutations et la parité ainsi que la ménopause. Cette étude montre l’importance de la prise en compte simultanée des facteurs de risque « non génétiques » et de la localisation des mutations dans les gènes BRCA1/2 dans l’estimation des risques de CS. Si nos résultats sont confirmés sur de plus larges données, cette étude pourrait aider ces femmes dans le choix du type de stratégie de surveillance ou de prévention le mieux adapté à leur situation. / Breast cancer is the most frequent and mortal cancer among French women. Mutations in BRCA1 and BRCA2 confer a high risk of breast cancer. The diversity of BRCA1 and BRCA2 mutations and of gyneco-obstetrical and life style risk factors may modify the magnitude of this risk. The objective of this study is to assess breast cancer risk taking into account gyneco-obstetrical and life style factors, location of mutation in BRCA1 and BRCA2 genes and there potential interaction.The data from 1337 women BRCA1 and BRCA2 mutation carriers of the French study GENEPSO have been analysed by using a weighted Cox regression model to take into account the bias of recruitment. Our results show an association between breast cancer risk and X-ray exposures, tobacco, body mass index, age at menarche, parity, interrupted pregnancies, oral contraceptives, menopausal status and hormone replacement therapy. We confirm the existence of a central low breast cancer risk region in BRCA1 (LR1 region between codons 374 and 1161) and in BRCA2 (LR2 region between codons 957 and 1827). In addition, a new high breast cancer risk region is found in the 3’ region of BRCA2 (HR2 region between codons 2546 and 2968). Moreover, our results suggest a variation in breast cancer risk associated with parity and menopausal status according to the location of the mutation in BRCA1 and BRCA2. Our findings show that, taking into account environmental and lifestyle modifiers, mutation position might be important for the clinical management of BRCA1 and BRCA2 mutation carriers and could also be helpful in understanding how BRCA1 and BRCA2 genes are involved in breast cancer.

Orientador: Luciana Patrícia Fernandes Abbade / Resumo: Introdução: A radioterapia tem demonstrado importante papel no controle loco-regional e à distância no tratamento de câncer de mama com aumento da taxa de sobrevida. Entretanto a radiodermite é um evento adverso com alta incidência nos tratamentos desta neoplasia. O objetivo deste estudo é avaliar a incidência de radiodermite e os fatores associados a sua gravidade e ao momento de ocorrência deste efeito colateral durante a radioterapia de mulheres com câncer de mama. Método: Estudo de coorte prospectivo e analítico, conduzido no Setor Técnico de Radioterapia do Hospital das Clínicas da Faculdade de Medicina de Botucatu – Unesp. Foram incluídas 117 pacientes com câncer de mama no período de fevereiro a setembro de 2017, com idade superior a 18 anos. As pacientes foram avaliadas na admissão e a cada cinco aplicações de radioterapia até o término do tratamento. Foram utilizados os testes de Qui-quadrado ou Exato de Fisher para as variáveis categóricas e Teste T para amostras independentes para as variáveis contínuas paramétricas e Kruskall-Wallis para as não paramétricas. Para a avaliação dos fatores de risco aplicou-se o modelo de regressão logística e modelo de regressão proporcional de Cox. Foram considerados significativos os valores de p menores que 0,05. Resultados: A taxa de incidência de radiodermite em topografia mamária foi de 98,2% e demonstrou que para cada ponto a mais do IMC aumenta-se em 14% a chance de ocorrência de radiodermite nos graus II a IV [OR = 1,14 (IC9... (Resumo completo, clicar acesso eletrônico abaixo) / Abstract: Introduction. The radiotherapy has an important role in locoregional and distant control of the neoplasia, increasing the disease-free and overall survival rates. However, the radiodermatitis is an adverse effect with high incidence in this treatment. The aim of this study is to evaluate the incidence of radiodermatitis in women with breast cancer, identify the factors associated with its severity, and determine the time this side effect occurs during radiotherapy. Methods. This was a prospective and analytical cohort study which was carried out in the Radiotherapy Sector of the Hospital das Clínicas de Botucatu, Botucatu Medical School (Brazil). One hundred seventeen patients included were women over 18 years of age who were diagnosed with breast cancer and admitted to the service from February to September 2017. The patients were evaluated at the time of inclusion in the study, before radiotherapy was started, and around every fifth application (with an evaluation window varying from three to seven applications) until the end of treatment. Chi-square or Fisher's exact tests were used for categorical variables and T-Test for independent samples for parametric continuous variables and Kruskall-Wallis for non-parametric variables. For the assessment of risk factors, the logistic regression and the Cox proportional regression model was used. P values lower than 0.05 were considered significant. Results. The incidence rate of radiodermatitis in breast area was 98.2% and it were ... (Complete abstract click electronic access below) / Mestre

This dissertation studies prediction and classification models for disease risk through semiparametric modeling and statistical learning. It consists of three parts. In the first part, we propose several survival models to analyze the Cooperative Huntington's Observational Research Trial (COHORT) study data accounting for the missing mutation status in relative participants (Kieburtz and Huntington Study Group, 1996a). Huntington's disease (HD) is a progressive neurodegenerative disorder caused by an expansion of cytosine-adenine-guanine (CAG) repeats at the IT15 gene. A CAG repeat number greater than or equal to 36 is defined as carrying the mutation and carriers will eventually show symptoms if not censored by other events. There is an inverse relationship between the age-at-onset of HD and the CAG repeat length; the greater the CAG expansion, the earlier the age-at-onset. Accurate estimation of age-at-onset based on CAG repeat length is important for genetic counseling and the design of clinical trials for HD. Participants in COHORT (denoted as probands) undergo a genetic test and their CAG repeat number is determined. Family members of the probands do not undergo the genetic test and their HD onset information is provided by probands. Several methods are proposed in the literature to model the age specific cumulative distribution function (CDF) of HD onset as a function of the CAG repeat length. However, none of the existing methods can be directly used to analyze COHORT proband and family data because family members' mutation status is not always known. In this work, we treat the presence or absence of an expanded CAG repeat in first-degree family members as missing data and use the expectation-maximization (EM) algorithm to carry out the maximum likelihood estimation of the COHORT proband and family data jointly. We perform simulation studies to examine finite sample performance of the proposed methods and apply these methods to estimate the CDF of HD age-at-onset from the COHORT proband and family combined data. Our results show a slightly lower estimated cumulative risk of HD with the combined data compared to using proband data alone. We then extend the approach to predict the cumulative risk of disease accommodating predictors with time-varying effects and outcomes subject to censoring. We model the time-specific effect through a nonparametric varying-coefficient function and handle censoring through self-consistency equations that redistribute the probability mass of censored outcomes to the right. The computational procedure is extremely convenient and can be implemented by standard software. We prove large sample properties of the proposed estimator and evaluate its finite sample performance through simulation studies. We apply the method to estimate the cumulative risk of developing HD from the mutation carriers in COHORT data and illustrate an inverse relationship between the cumulative risk of HD and the length of CAG repeats at the IT15 gene. In the second part of the dissertation, we develop methods to accurately predict whether pre-symptomatic individuals are at risk of a disease based on their various marker profiles, which offers an opportunity for early intervention well before definitive clinical diagnosis. For many diseases, existing clinical literature may suggest the risk of disease varies with some markers of biological and etiological importance, for example age. To identify effective prediction rules using nonparametric decision functions, standard statistical learning approaches treat markers with clear biological importance (e.g., age) and other markers without prior knowledge on disease etiology interchangeably as input variables. Therefore, these approaches may be inadequate in singling out and preserving the effects from the biologically important variables, especially in the presence of potential noise markers. Using age as an example of a salient marker to receive special care in the analysis, we propose a local smoothing large margin classifier implemented with support vector machine to construct effective age-dependent classification rules. The method adaptively adjusts age effect and separately tunes age and other markers to achieve optimal performance. We derive the asymptotic risk bound of the local smoothing support vector machine, and perform extensive simulation studies to compare with standard approaches. We apply the proposed method to two studies of premanifest HD subjects and controls to construct age-sensitive predictive scores for the risk of HD and risk of receiving HD diagnosis during the study period. In the third part of the dissertation, we develop a novel statistical learning method for longitudinal data. Predicting disease risk and progression is one of the main goals in many clinical studies. Cohort studies on the natural history and etiology of chronic diseases span years and data are collected at multiple visits. Although kernel-based statistical learning methods are proven to be powerful for a wide range of disease prediction problems, these methods are only well studied for independent data but not for longitudinal data. It is thus important to develop time-sensitive prediction rules that make use of the longitudinal nature of the data. We develop a statistical learning method for longitudinal data by introducing subject-specific long-term and short-term latent effects through designed kernels to account for within-subject correlation of longitudinal measurements. Since the presence of multiple sources of data is increasingly common, we embed our method in a multiple kernel learning framework and propose a regularized multiple kernel statistical learning with random effects to construct effective nonparametric prediction rules. Our method allows easy integration of various heterogeneous data sources and takes advantage of correlation among longitudinal measures to increase prediction power. We use different kernels for each data source taking advantage of distinctive feature of data modality, and then optimally combine data across modalities. We apply the developed methods to two large epidemiological studies, one on Huntington's disease and the other on Alzhemeier's Disease (Alzhemeier's Disease Neuroimaging Initiative, ADNI) where we explore a unique opportunity to combine imaging and genetic data to predict the conversion from mild cognitive impairment to dementia, and show a substantial gain in performance while accounting for the longitudinal feature of data.

Introdução: Nascimento de pré-termo é o que ocorre com menos de 37 semanas completas de gestação, e é tido como um dos principais fatores de risco de morbimortalidade neonatal. Sua etiologia é complexa e envolve inúmeros fatores. Objetivo: Identificar os fatores de risco associados aos nascimentos pré-termo. Método: Estudo populacional tipo caso-controle. Foram estudados nascidos vivos hospitalares de mães residentes em Londrina (PR) entre junho de 2006 e março de 2007. Os casos foram 328 nascimentos com idade gestacional inferior a 37 semanas (pré-termo) e os controles uma amostra (369) representativa de nascimentos com 37 semanas ou mais de idade gestacional. Os dados foram obtidos por meio de entrevistas com as mães no hospital e dos prontuários da mãe e do recém nascido. As variáveis estudadas foram agrupadas em cinco blocos, representando diferentes níveis de hierarquia: características socioeconômicas; características pré-concepcionais e historia reprodutiva materna; condições da gestação; agravos maternos na gestação e características fetais. Foi realizada análise de regressão logística múltipla hierarquizada. Resultados: No modelo final foi identificada associação estatisticamente significante (P<0,05) ao nascimento pré-termo para as seguintes categorias de variáveis: no bloco 1 o local de moradia em favela (OR=1,80; IC 95%:1,02-3,19) e a baixa idade (16 a 29 anos) do chefe da família(1,57; 1,15-2,13); no bloco 2 as mães com IMC<19kg/m2-magras (2,12; 1,40-3,21) ou IMC >= 30 kg/m2 - obesas (1,96; 1,04-3,70), com nascimento anterior pré-termo (3,04; 1,83-5,05) e que realizaram tratamento para engravidar (8,28; 2,70-25,41); no bloco 3 as mães com companheiro há menos de 2 anos (1,44; 1,03-2,00), que relataram preocupações na gestação (1,52; 1,05-2,21), que consumiram bebida alcoólica semanalmente (2,52; 1,19-5,36) e que receberam assistência pré-natal inadequada (3,57; 1,48-8,58), praticar caminhada na gestação foi um fator protetor (0,48; 0,33-0,70); no bloco 4 as mães que apresentaram sangramento (5,19; 2,54-10,60), infecção do trato genital (2,96; 1,09-7,99), alteração do volume amniótico (5,82; 2,32-14,56), hipertensão arterial com eclampsia (8,67; 4,09-18,37) e sem eclampsia (1,9; 1,01-3,61) e internação na gestação (5,55; 2,86-10,77); no bloco 5 apenas gestação gemelar foi fortemente identificada como risco (20,10; 4,44-90,99). Conclusão: Condições socioeconômicas desfavoráveis juntamente com características biológicas maternas, intercorrências da gestação e condições psicossociais adversas constituem-se em risco para nascimentos pré-termo. O aprimoramento da qualidade da atenção pré-natal, incluindo a identificação destes fatores na gestação pode reduzir a prematuridade. / Introduction: Pre-term birth happens when the fetus is born before 37 weeks of pregnancy and is considered one of the main risk factors of neonatal morbidity and mortality. It has a complex etiology and a wide scope. Objective: Identify the risk factors associated with pre-term birth. Method: Population-based case-control study. Children born alive in Londrina hospitals from June 2006 to March 2007 were studied. Cases were analyzed from 328 births with pregnancy length below 37 weeks (pre-term) and checks were studied from a birth representative sample (369) with pregnancy length above 37 weeks or more. Data were obtained from interviews carried out with mothers at the hospital and mothers and newborns medical reports. Variables studied were grouped in five groups, representing different hierarchical levels: socioeconomic characteristics; pre-conceptive characteristics and reproduction history; pregnancy conditions; pregnancy problems and fetus characteristics. A multiple hierarchical logistics regression analysis was carried out. Results: There was a significant relation (P<0.05) between pre-term birth and the following variables: group 1 - place - shanty towns(OR=1.80; CI 95% 1.02,3.19); and head of the family aged between16 and 29 years old (OR=1.57; CI 95% 1.15,2.13); group 2 - mothers presenting a BMI < 19 kg/m2 - thin mothers (OR=2.12; CI 95% 1.40,3.21) or BMI >=30 kg/m2 - obese mothers (OR=1.96; CI 95% 1.04,3.70), history of a pre-term birth (OR=3.04; CI 95% 1.83,5.05) and among women undertaking fertility treatment (OR=8.28; CI 95% 2.70,25.41); group 3 - mothers who have been in a relationship with the same partner for less than two years (OR=1.44; CI 95% 1.03,2.00) , stressed mothers (OR=1.52; CI 95% 1.05,2.21), mothers who consumed alcoholic beverages weekly (OR=2.52; CI 95% 1.19,5.36) and mothers who had not taken appropriate prenatal care (OR=3.57; 1.48-8.58), regular walks during pregnancy was a protector factor (OR=0.48; CI 95% 0.33-0.70); group 4 - mothers with bleeding onset (OR=5.19; CI 95% 2.54,10.60); genital tract infection (OR=2.96; CI 95% 1.09,7.99), amniotic volume alterations (OR=5.82; CI 95% 2.32,14.56), high blood pressure with (OR=8.67; CI 95% 4.09,18.37) and without eclampsia (OR=1.91;CI 95% 1.01,3.61) and hospital admission (OR=5.55; CI 95% 2.86,10.77); group 5 - twin pregnancy was strongly associated with risk (OR=20.10; CI 95% 4.44,90.99). Conclusion: An unfavorable socio-economic condition added to some maternal biological characteristics, pregnancy-related health problems and psychosocial conditions constitute a high risk for pre-term births. An improvement of the pre-natal quality service and the identification of the factors listed above during pregnancy may reduce premature birth.

O objetivo do estudo foi caracterizar a situação epidemiológica da brucelose e tuberculose bovinas na região três do Estado de São Paulo. Assim, 260 rebanhos com atividade reprodutiva foram aleatoriamente selecionados e, dentro de cada um deles, 10 ou 15 fêmeas bovinas com idade igual ou superior a 24 meses foram aleatoriamente escolhidas e testadas para o diagnóstico da brucelose. Foi utilizado um protocolo de testes em série, tendo o teste do antígeno acidificado tamponado como método de triagem e o teste de fixação de complemento como confirmatório. Para o diagnóstico da tuberculose foi utilizado o teste tuberculínico cervical comparativo em 20 ou 40 fêmeas bovinas com idade igual ou superior a 24 meses. A prevalência de focos, tanto de brucelose quanto de tuberculose, foi de 7,3% [4,44%; 11,17%]; a prevalência de animais positivos foi de 1,37% [0,91%; 2,00%] para brucelose e 1,05% [0,73%; 1,46%] para tuberculose. Em cada propriedade foi aplicado um questionário epidemiológico para individualizar possíveis fatores de risco para as duas doenças. A variável associada à condição de foco de brucelose foi a propriedade ter 23 ou mais bovinos com idade igual ou superior a 24 meses (OR=5,0 [1,83; 13,7]); a aquisição de bovinos emergiu como fator de risco para a tuberculose (OR= 4,46 [1,55; 12,78]) . / The purpose of this study is to propound bovine brucellosis and tuberculosis epidemiological situation at region 3 in São Paulo. Thus, 260 reproduce activity cattle were been randomly chosen and, among it, 10 or 15 female bovine at the same age or up 24 months were randomly chosen and tested for the brucellosis diagnosis. Test serial protocol was made using, having the test buffered acidified as a screening method and complement fixation test as a confirmatory. Comparative cervical tuberculin test was made using in 20 or 40 female bovine at the same age or up 24 months. The prevalence of focus at both brucellosis and tuberculosis was been7,3% [4,44%; 11,17%]; the prevalence of positive animals was been 1,37% [0,91%; 2,00%] for brucellosis and 1,05% [0,73%; 1,46%] for tuberculosis. A epidemiological questionnaire has been applied on each property as a purpose to individualize possible risk factors at both disease. The variable associated with tuberculosis condition was the property has 23 or more female bovine at the same age or up 24 months (OR=5,0 [1,83; 13,7]); cattle purchase has emerged as a risk factor for tuberculosis (OR= 4,46 [1,55; 12,78]) .

Introdução: Estudos prévios analisaram a associação entre fatores de risco cardiovascular (FRCV) associados ao diagnóstico de demência vascular (DV) provável ou possível. No entanto, não foram encontrados estudos que analisassem a associação entre FRCV e a ocorrência de DV definitiva. Dessa maneira, ainda permanece obscura a associação entre os FRCV e a ocorrência de DV definitiva, ou seja, aquela diagnosticada por meio do exame neuropatológico, no qual se apresenta como padrão ouro. Objetivo: Avaliar a associação entre os FRCV e a ocorrência de DV definitiva, pura e mista. Método: Por meio de um estudo transversal foram analisados 707 casos pertencentes à casuística do Banco de Encéfalos Humanos do Grupo de Estudos em Envelhecimento (BEHGEEC) da FMUSP, que respeitaram os critérios de inclusão. A existência de fatores de risco cardiovascular em vida (Hipertensão Arterial, Diabetes Mellitus, Dislipidemia, Tabagismo, Etilismo, Obesidade e Sedentarismo), reportada por um informante com convivido minimamente semanal durante a autópsia, foi associada ao diagnóstico neuropatológico de demência vascular emitido por um neuropatologista. Modelos de regressão logística (sem e com ajuste para sexo, idade e raça) foram construídos para testar a associação entre os FRCV e o diagnóstico de DV, DV pura e DV mista. Foi testada a capacidade preditiva dos fatores que se mostraram preditores de DV por meio da Curva ROC. Resultados: O sedentarismo foi um preditor independente de DV (OR 1,943; IC95% 1,198 3,151; p= 0,007) e DV pura (OR 3,148; IC95% 1,428 6,941;p= 0,004). A HAS foi um preditor independente de DV mista (OR 2,240; IC95% 1,216 4,126; p= 0.01). O sedentarismo não apresentou boa capacidade preditiva para a DV e DV Pura (AUC = 0,380 e 0,337, respectivamente), assim como a HAS para a DV Mista (AUC = 0,459). Conclusões: Dentre os FRCV o sedentarismo e a HAS foram os que se associaram a um aumento no risco de DV. / Introduction: Previous studies have analyzed the association between cardiovascular risk factors (CVRF) associated with the diagnosis (probable or possible) of vascular dementia (VaD). However, there are no studies that have analyzed the association between CVRF and the occurrence of definitive VaD. The association between CVRF and the occurrence of definite VaD, neuropathologically defined and considered as gold-standard, remains obscure. Objectives: To evaluate the association between CVRF and the occurrence of definitive VaD, pure and mixed. Methods: This is a cross-sectional study which evaluated 707 cases belonging to the Bain Bank of the Brazilian Aging Brain Study Group (BBBABSG) of FMUSP, respecting the inclusion criteria. The history of existence of cardiovascular risk factors in life (hypertension, diabetes mellitus, dyslipidemia, smoking, alcoholism, obesity, and sedentarism) reported by a knowledgeable next-of-kin, with at least weekly contact with the deceased, was associated with the neuropathological diagnosis of vascular dementia reported by a neuropathologist after the autopsy exam. Logistic regression models (with and without adjustment for sex, age and race) were tested to show the association between CVRF and the diagnosis of VaD, pure Vad and mixed VaD. It was also tested the predictive capacity of the factors that proved to be predictors of VaD through the ROC Curve. Results: Sedentary lifestyle was an independent predictor of VaD (OR 1,943, CI 95% 1,198 - 3,151, p = 0.007) and of Pure VaD (OR 3,148, 95% CI, 1.428 - 6.941, p = 0.004). Hypertension was an independent predictor of Mixed VaD (OR 2,240, 95% CI 1,216 - 4,126, p = 0.01). Sedentary lifestyle did not present good predictive capacity for VaD and Pure VaD (AUC = 0.380 and 0.337, respectively), as Hypertension for Mixed DV did not either (AUC = 0.459). Conclusions: Among the CVRF, sedentarism and hypertension were those associated with an increase VaD risk.