Tumas Natalia, Niclis Camila, Aballay Laura R, Osella Alberto R, Díaz María del Pilar
Statistics and Biostatistics Unit, School of Nutrition, Faculty of Medical Sciences, National University of Córdoba, Córdoba, Argentina,
Eur J Nutr. 2014;53(2):557-66. doi: 10.1007/s00394-013-0564-0. Epub 2013 Aug 2.
Several studies have shown the effect of dietary patterns on breast cancer risk, but none has been conducted in Argentina. The aim of this study was to extract dietary patterns from Food Frequency Questioner, to estimate their effect on breast cancer occurrence while taking into account aggregation factors (family history of breast cancer) and to explore the sensitivity of the estimates to changes in the assumptions.
A principal component exploratory factor analysis was applied to identify dietary patterns, which were then included as covariates in a multilevel logistic regression. Family history of BC was considered as a clustering variable. A multiple probabilistic sensitivity analysis was also performed.
The study included 100 cases and 294 controls. Four dietary patterns were identified. Traditional (fat meats, bakery products, and vegetable oil and mayonnaise) (OR III tertile vs I 3.13, 95% CI 2.58-3.78), Rural (processed meat) (OR III tertile vs I 2.02, 95% CI 1.21-3.37) and Starchy (refined grains) (OR III tertile vs I 1.82, 95 % CI 1.18-2.79) dietary patterns were positively associated with BC risk, whereas the Prudent pattern (fruit and non-starchy vegetables) (OR III tertile vs I 0.56, 95% CI 0.41-0.77) showed a protective effect. For Traditional pattern, the median bias-adjusted ORs (3.52) were higher than the conventional (2.76).
While the Prudent pattern was associated with a reduced risk of BC, Traditional, Rural and Starchy patterns showed a promoting effect. Despite the threats to validity, the nature of associations was not strongly affected.
多项研究表明饮食模式对乳腺癌风险有影响，但阿根廷尚未开展此类研究。本研究旨在从食物频率问卷中提取饮食模式，在考虑聚集因素（乳腺癌家族史）的情况下估计其对乳腺癌发生的影响，并探讨估计值对假设变化的敏感性。
应用主成分探索性因子分析来识别饮食模式，然后将其作为协变量纳入多级逻辑回归。乳腺癌家族史被视为聚类变量。还进行了多重概率敏感性分析。
该研究纳入了100例病例和294例对照。识别出四种饮食模式。传统模式（肥肉、烘焙食品、植物油和蛋黄酱）（第三三分位数与第一三分位数相比的优势比为3.13，95%置信区间为2.58 - 3.78）、农村模式（加工肉类）（第三三分位数与第一三分位数相比的优势比为2.02，95%置信区间为1.21 - 3.37）和淀粉类模式（精制谷物）（第三三分位数与第一三分位数相比的优势比为1.82，95%置信区间为1.18 - 2.79）与乳腺癌风险呈正相关，而谨慎模式（水果和非淀粉类蔬菜）（第三三分位数与第一三分位数相比的优势比为0.56，95%置信区间为0.41 - 0.77）显示出保护作用。对于传统模式，中位数偏差调整后的优势比（3.52）高于常规值（2.76）。
虽然谨慎模式与乳腺癌风险降低相关，但传统模式、农村模式和淀粉类模式显示出促进作用。尽管存在有效性威胁，但关联的性质并未受到强烈影响。
