Os dados utilizados nesta tese fazem parte do estudo "Doenças cardiovasculares ateroscleróticas, dislipidemias, hipertensão, obesidade e diabetes melito em população da área Metropolitana de São Paulo (SP)." Corresponde a um estudo transversal em uma amostra constituída de 1.045 adultos, civis, de ambos os sexos, de 20 anos ou mais, não institucionalizados, residentes no Município de Cotia, situado a sudoeste da área Metropolitana de São Paulo. O levantamento dos dados foi realizado através de entrevista direta, na forma de questionário padronizado, composto de inquérito sócio-demográfico, inquérito clínico, medidas antropométricas, bioquímico, estilo de vida e inquérito dietético. Este último inquérito foi baseado na história alimentar do individuo. A partir de uma lista de alimentos, construída com dados de um estudo piloto realizado na região, fez-se o registro da freqüência de consumo de cada alimento durante o último ano. Os alimentos identificados foram classificados em uma das sete categorias de freqüências de consumo; a partir das quais foram determinados os: escore I (formado pelos alimentos reconhecidos como de risco das doenças cardiovasculares) e o escore II (formado por alimentos como benéficos no risco das doenças cardiovasculares). A partir da freqüência de consumo testaram-se relações entre consumo alimentar e perfil lipídico da população estudada. Os escores foram analisados segundo características demográficas da população (sexo, idade e nível de instrução). Também, procurou-se verificar possíveis relações entre os escores de consumo e as frações de lipídeos séricos, C-TOTAL, LDL- C e HDL-C, através do coeficiente de correlação de Spearman, e de modelos de regressão linear multivariada (stepwise forward). Utilizou-se o nível de significância de a = 5%. Aspectos relativos ao consumo alimentar mostram que os cereais refinados e derivados e o feijão estiveram presente diariamente na alimentação de quase o total da população. Quase um terço da população não consome hortaliças e dois terços não consumiram frutas. O padrão de consumo foi constituído por cereais refinados, gordura de origem vegetal, café, açúcar e feijão. A distribuição do escore I e II não apresentou associação significativa do ponto de vista estatístico segundo sexo, faixa etária e nível de instrução, com exceção do sexo feminino com idade de 50 a mais que apresentou diferença estatisticamente significativa entre quintis do escore II e nível de instrução. Observou-se um aumento gradativo de C-TOTAL e LDL-C a partir do primeiro ao quinto quintil de consumo dos alimentos que compõem o escore I. Os indivíduos inseridos nos quintis superiores do escore II, apresentaram uma média de C-TOTAL e LDL-C mais baixo do que aqueles enquadrados no quintil inferior. Constatou-se correlação positiva e significativa e negativa e significativa entre os escore I, e escore II, respectivamente, com níveis de C- TOTAL e LDL-C. Todos os alimentos de origem animal apresentaram coeficiente de correlação de Spearman positivo e significativo com C-TOTAL e LDL-C, com exceção da manteiga. O consumo de frutas e hortaliças e carne de peixe apresentou coeficiente de correlação de Spearman inverso e significativo com C-TOTAL e LDL-C. HDL-C apresentou correlação positiva e significativa com o consumo de bebidas alcoólicas e, inversa e significativa com o consumo de feijão. Nos modelos de regressão linear multivariada elaborados para C-TOTAL e LDL-C o escore I e o consumo de carnes (bovina, suina, aves, víscera e carnes processadas), leite e derivados, e ovos, correlacionou-se positiva e significativamente com esses constituintes sangüíneos. O escore II e o consumo de frutas e hortaliças mostraram correlação inversa e significativa. O consumo de bebidas alcoólicas se correlacionou positiva e significativamente com a fração HDL-C. Entre os três modelos elaborados para as frações de lipídeos séricos (C-TOTAL, LDL-C e HDL-C), após o ajuste por variáveis alimentares e não alimentares, o modelo de C-TOTAL apresentou maior capacidade explicativa, R2= 24. Conclue-se que Cotia tem uma população com altas proporções de fatores de risco bem definidos na etiologia das dislipidemias e, conseqüentemente, fatores de risco para DCV.

The data used in this thesis is part of the study "Ateroschlerotic cardiovascular disease, lipemic disorders, hypertension, obesity and diabetes mellitus in the population of a metropolitan area of southeastern Brazil". These observations were collected during a cross-sectional exarnination, and a non-institutionalized population was involved in the sample. The survey was administered to a representative sample of adults between 20 years of age and over, from Cotia City, São Paulo Metropolitan area of southeastern Brazil. The questionnaire consisted of differents sections: demographic, clinical, anthropometric, biochemical, lifestyle and, a food frequency section. Data on food frequency consumption, serum lipids, and other covariates were avaliable for 1045 adults out of a total of 1328 interviewed during the survey. A twelve months retrospective Simple Food Frequency Questionnaire, open end, was used to determine food-consumption frequency. Each participant was interviewed by a trained staff member or a dietitian using a previous evaluated instrument in a pilot study. The questionnaire was representative of the usual diet of the individuais. For each of the food items, frequency of consumption was coded into one to seven categories: from never to every day. Food consumption was examined in two differents ways: by the score of consumption, which was obtained by grouping food according to their compositions into two large groups, score I (recognized as risk food for cardiovascular risk) and escore II, (recognized as healthy food for cardiovascular risk) and, by the different food items. The food patterns population caracterized by the score of consumption was examined by gender, age and level of education. The correlation between serum lipids, TOTAL-C, LDL-C, HDL-C and, scores of food consumpion was also examined. Values from the Spearman-correlation analyses were used to provide an estimate of the linear correlation between individual food items consumption and serum lipid concentrations. The association between food consumption frequency and serum blood lipoproteins levels among the study population was analyzed with multiple linear regression models. The measurements of serum lipoproteins (TOTAL-C, LDL-C, and HDL-C) levels were the dependent variables (three models), and entered into each linear regression model with age, gender, other non-dietary and dietary covariate as independent variables. The food consumption frequency for each food was added to the models as a continuous variable. Modelling step-wise techniques were used to enter the covariates into the linear models. Significance levels was defined as a = 5%. In order to examine the quality of the diet, food frequency consumption was observed. The higher consumption was bread, rice and other refined products and beans. Almost 1/3 of the population did not consume vegetables and 2/3 did not consume fruits on the daily bases. The food pattern showed by the population was composed by: refined cereais, vegetables fat, coffee, sugar and beans. Difference of consumption among sex, age groups and educational level were not observed. Increasing mean level of TOTAL-C and LDL-C was observed from the lowest to the highest quintile of score I, when comparing to the score II, decreasing mean level of TOTAL-C and LDL-C was observed from the lowest to the highest quintile. Values from the Spearman-correlation analyses were used to provide an estimate of the linear association between individual food items consumption frequency and serum lipid concentration. In the Spearman-correlation analyses, food consumption was treated as a continuos variable. Spearman correlation coefficients between TOTAL-C, LDL-C and score I was positively significant and the score II showed to be negatively significant. All animal products showed to be positively and significant correlated with TOTAL-C, LDL-C levels when Spearman correlation coefficients was observed by food items, only butter didn't showed this correlation. The consumption for vegetables and fish showed inverse and significant correlation with TOTAL-C and LDL-C. Positive and significant correlation was observed between HDL-C levels and alcohol consumption, inverse correlation was observed for beans consumption. The results of the linear regression analyses between serum TOTAL-C and LDL-C levels and food frequency consumption, after multivariate adjusted, showed significant positive relationship between score I, meat consumption (red meat, chicken, pork, liver and others meats from internal organs), dairy foods, eggs and, showed significant and inverse relationship between score II, fruits and vegetables consumption. No significant association was found for the consumption of alcohol, with and without adjusting for the potencial confounders. For HDL-C alcohol consumption was also positively significantly associated after multivariate adjusted. R2 goodness-of-fit values showed that the TOTAL-C fitted model, explain 24% of the variation in the TOTAL-C concentration. The other models explained less variations. Cotia population showed low level of education, low income and high proportions of lipemic disorders risk factors.