Swerdlow Anthony J, De Stavola Bianca L, Floderus Birgitta, Holm Niels V, Kaprio Jaakko, Verkasalo Pia K, Mack Thomas
Section of Epidemiology, Institute of Cancer Research, Sutton, UK.
J Natl Cancer Inst. 2002 Aug 21;94(16):1238-46. doi: 10.1093/jnci/94.16.1238.
Breast cancer etiology in women may relate to exposures early in life as well as in adulthood, but it has been difficult to gain information on childhood variables, and evidence on their effects is very limited. Comparison of risk factor levels between affected probands and unaffected twins in twin pairs can provide a unique method to investigate risk factors that act in early life.
We conducted four population-based case-control studies of breast cancer risk in twins in Denmark, England and Wales, Finland, and Sweden and pooled the data from these studies. The case patients were 400 women with incident breast cancer before the age of 50 years, and the control subjects were their twin sisters who had not had breast cancer by that time. Data on risk factors (e.g., demographic and anthropomorphic variables, reproductive history, and family history) were collected by interview and by a mailed questionnaire and were analyzed by calculating matched odds ratios [ORs]. All statistical tests were two-sided.
The risk of breast cancer was increased for women who were less obese (OR = 1.44, 95% confidence interval [CI] = 1.08 to 1.91) or taller (OR = 1.27, 95% CI = 0.95 to 1.70) than their co-twin at age 10 years, for women who developed breasts earlier than their co-twin (OR = 1.53, 95% CI = 1.14 to 2.06), and for women who had a smaller waist-to-hip ratio at age 20 years than their co-twin (OR = 1.79, 95% CI = 1.00 to 3.21). Analysis according to family history of breast cancer showed that the associations of childhood height and weight with risk of breast cancer were only apparent in women without a family history of breast cancer.
Childhood growth before puberty may affect the risk of premenopausal breast cancer, at least in women without a family history of breast cancer. The distribution of body fat in young adulthood may also be related to breast cancer risk.
女性乳腺癌的病因可能与生命早期以及成年期的暴露因素有关，但获取童年时期变量的信息一直很困难，且关于其影响的证据非常有限。比较患病先证者与未患病双胞胎之间的危险因素水平，可为研究生命早期起作用的危险因素提供一种独特方法。
我们在丹麦、英格兰和威尔士、芬兰以及瑞典对双胞胎中的乳腺癌风险开展了四项基于人群的病例对照研究，并汇总了这些研究的数据。病例组为400名50岁之前患原发性乳腺癌的女性，对照组为当时未患乳腺癌的她们的双胞胎姐妹。通过访谈和邮寄问卷收集危险因素（如人口统计学和人体测量学变量、生殖史和家族史）数据，并通过计算配对比值比[OR]进行分析。所有统计检验均为双侧检验。
与10岁时比其双胞胎姐妹瘦（OR = 1.44，95%置信区间[CI] = 1.08至1.91）或高（OR = 1.27，95%CI = 0.95至1.70）的女性、乳房发育比其双胞胎姐妹早的女性（OR = 1.53，95%CI = 1.14至2.06）以及20岁时腰臀比比其双胞胎姐妹小的女性（OR = 1.79，95%CI = 1.00至3.21）相比，她们患乳腺癌的风险增加。根据乳腺癌家族史进行的分析表明，童年时期的身高和体重与乳腺癌风险之间的关联仅在无乳腺癌家族史的女性中明显。
青春期前的儿童生长可能会影响绝经前乳腺癌的风险，至少在无乳腺癌家族史的女性中如此。成年早期身体脂肪的分布也可能与乳腺癌风险有关。
