Hsieh C C, Trichopoulos D, Katsouyanni K, Yuasa S
Department of Epidemiology, Harvard School of Public Health, Boston, MA 02115.
Int J Cancer. 1990 Nov 15;46(5):796-800. doi: 10.1002/ijc.2910460508.
The importance of age at menarche, age at menopause, height, and obesity as risk factors for breast cancer, and the possible interactions among these factors in breast cancer causation were investigated in a data set collected in the late 1960's, in an international multicenter case-control study. Multiple logistic regression procedures were used to model data from 3,993 breast cancer cases and 11,783 controls from 7 study centers representing the range of international variation of breast cancer incidence. Height and obesity (measured through the weight/height2 index) were independent risk factors for breast cancer among post-menopausal but not pre-menopausal women; post-menopausal women taller by 10 cm had a 12% higher risk of breast cancer (95% confidence interval, CI, 3-21%) and post-menopausal women of average height (say 158 cm) had an 11% higher risk of breast cancer (CI 7-16%) when they were heavier by 10 kg (and, therefore, more obese by 4 kg/m2). Age at menarche was a risk factor among both pre-menopausal and post-menopausal women, a delay of 2 years corresponding to a 10% reduction in breast cancer risk (CI 6-15%). Age at menopause was also a breast cancer risk factor, women with menopause at each 5 year age difference having a 17% higher risk of breast cancer (CI 11-22%). There is evidence of an interaction (deviation from the logistic regression-postulated multiplicativity) between obesity and age at menarche, implying that the protective effect of late menarche may not apply to overweight women or that late menarche may become detrimental in obese women. The estimated relative risk coefficients, when applied to average risk factor levels observed among control women, can explain only a small fraction of the difference in breast cancer incidence between Boston and Tokyo.
在一项于20世纪60年代末收集数据集的国际多中心病例对照研究中，对初潮年龄、绝经年龄、身高和肥胖作为乳腺癌风险因素的重要性，以及这些因素在乳腺癌病因中的可能相互作用进行了调查。使用多元逻辑回归程序对来自7个研究中心的3993例乳腺癌病例和11783例对照的数据进行建模，这些研究中心代表了乳腺癌发病率的国际变化范围。身高和肥胖（通过体重/身高²指数衡量）是绝经后而非绝经前女性患乳腺癌的独立风险因素；绝经后女性身高每增加10厘米，患乳腺癌的风险就高12%（95%置信区间，CI，3%-21%），平均身高（比如158厘米）的绝经后女性体重增加10千克（因此，肥胖程度增加4千克/平方米）时，患乳腺癌的风险高11%（CI 7%-16%）。初潮年龄在绝经前和绝经后女性中都是一个风险因素，初潮延迟2年相当于乳腺癌风险降低10%（CI 6%-15%）。绝经年龄也是乳腺癌的一个风险因素，每相差5岁绝经的女性患乳腺癌的风险高17%（CI 11%-22%）。有证据表明肥胖与初潮年龄之间存在相互作用（偏离逻辑回归假定的可乘性），这意味着初潮晚的保护作用可能不适用于超重女性，或者初潮晚在肥胖女性中可能变得有害。当将估计的相对风险系数应用于对照女性中观察到的平均风险因素水平时，只能解释波士顿和东京乳腺癌发病率差异的一小部分。
