【原】糖尿病女性的乳腺癌与死亡风险

　　编者按：全球乳腺癌和糖尿病的发病率呈上升趋势，深入了解这两种疾病之间的相关性，对于改善治疗和生存结局可能至关重要。本研究根据芬兰全国数据登记库女性队列研究，发现糖尿病（被诊断于乳腺癌之前）为乳腺癌死亡的风险因素，无论参加乳腺钼靶筛查与否，均与乳腺癌的合并症、肿瘤分期和组织学分级密切相关。当糖尿病被诊断于乳腺癌之后，该风险相关性变弱。这些研究结果表明，慢性糖尿病或可预测致命乳腺癌。

　　2018年1月23日，国际抗癌联盟《国际癌症杂志》在线发表芬兰坦佩雷大学医院、国家卫生福利研究所、坦佩雷大学、美国约翰霍普金斯大学布隆博格公共卫生学院的芬兰全国女性乳腺癌患者队列研究报告，调查了糖尿病与乳腺癌诊断时分期和后续生存的相关性。

　　该研究通过芬兰癌症登记库（覆盖99%以上的芬兰癌症患者）获得1995～2013年所有新诊断为乳腺癌并且记录死亡原因和日期的患者资料共7万3170例；通过国家卫生福利研究所普查登记库（芬兰政府自1980年对全国50～69岁女性开展每两年一次的乳腺钼靶筛查，覆盖率达90.9%）获得乳腺钼靶筛查的参加情况；通过全国医疗登记库获得1995～2013年记录的糖尿病诊断和背景情况信息，并与全国处方登记库用药数据合并；通过分类评定逻辑回归，校正乳腺钼靶筛查和乳腺癌诊断时年龄的影响，对乳腺癌诊断时分期较晚的风险进行定量分析；通过生存曲线比例风险回归，对总生存和乳腺癌生存进行定量分析。上述分析均对年龄、背景情况、乳腺钼靶筛查的影响进行了校正，生存分析还对肿瘤分期、组织学分级、基础治疗的影响进行了校正。

　　结果发现，经过中位5.8年的随访：

• 总死亡患者2万2520例（30.8%）

• 乳腺癌死亡1万0900例（14.9%）

• 糖尿病患者1万1676例（16.0%）

　　无论糖尿病与否，乳腺钼靶筛查率相似。

　　乳腺癌诊断时，糖尿病与非糖尿病相比：

• 局部晚期乳腺癌多1.26倍（95%置信区间：1.18～1.35）

• 远处转移乳腺癌多1.59倍（95%置信区间：1.44～1.75）

　　乳腺癌诊断后，糖尿病与非糖尿病相比：

• 乳腺癌死亡风险高1.36倍（95%置信区间：1.27～1.46）

• 内分泌治疗率较低（34.5%比38.4%，P<0.001）

　　患糖尿病越早，乳腺癌死亡风险越高：

• 先有糖尿病：风险高1.36倍（95%置信区间：1.25～1.47）

• 后有糖尿病：风险高1.18倍（95%置信区间：1.10～1.26）

　　患糖尿病越久，乳腺癌死亡风险越高：

• 先有糖尿病≥5年：风险高1.42倍（95%置信区间：1.28～1.57）

• 先有糖尿病<5年：风险高1.32倍（95%置信区间：1.20～1.45）

　　因此，糖尿病女性与非糖尿病女性相比，被诊断为乳腺癌的风险较高、分期较晚，即使考虑分期和筛查的影响之后，乳腺癌死亡风险仍然较高。如果糖尿病诊断早于乳腺癌诊断，那么长期糖尿病与致命乳腺癌风险的相关性最强。该研究结果支持将糖尿病作为致命乳腺癌的风险因素。

　　不过，虽然该研究的样本量足够大、采样期足够长，但是缺少激素补充疗法、激素受体状态、肿瘤分级、体重指数、肥胖或超重、社会经济地位、吸烟等生活方式的影响因素分析，而且相关性也不等于必然性，芬兰女性更无法完全代表所有女性，以上研究结果仅供专业人士参考，切勿片面理解。

Int J Cancer. 2018 Jan 23. [Epub ahead of print]

Breast cancer extent and survival among diabetic women in a Finnish nationwide cohort study.

Murto MO, Artama M, Pukkala E, Visvanathan K, Murtola TJ.

Tampere University Hospital, Tampere, Finland; National Institute for Health and Welfare, Tampere, Finland; University of Tampere, Tampere, Finland; Johns Hopkins Bloomberg School of Public Health, Baltimore, MD.

Breast cancer incidence and diabetes mellitus are on the rise globally. A better understanding of associations between the two diseases could prove critical to improving treatment and survival outcomes. Here, in a nationwide register-based cohort study of women in Finland, diabetes mellitus was found to be a risk factor for breast cancer death, regardless of mammography screening participation, closely associated comorbidities, and tumor extent and histology. The risk association was weaker when diabetes was diagnosed after breast cancer. The findings suggest that chronic diabetes could be predictive of fatal breast cancer.

Breast cancer (BC) and diabetes mellitus (DM) are major health problems. We examined the association between DM and BC stage at diagnosis and subsequent survival in a Finnish cohort of female BC patients. All BC cases (N=73,170) diagnosed in 1995-2013 with dates and causes of death were identified from the Finnish Cancer Registry. Participation in organized mammography screening was obtained from Mass Inspection Registry. Information on DM diagnoses and background conditions recorded during 1995-2013 were obtained from national Care Register for Health Care and merged to data on medication use from the national Prescription Register. Logistic regression with adjustment for mammography screening and age at BC diagnosis was used to evaluate the risk of advanced stage BC at diagnosis. Cox regression was used to evaluate overall and BC survival. Analyses were adjusted for age, background conditions and mammography screening. Survival analyses were further adjusted for tumor extent, histology and primary treatment. Of the cohort 11,676 (16.0%) had DM. Screening participation did not differ by diabetes. Compared to non-diabetic women, diabetics had more often locally advanced (odds ratio, OR 1.26; 95% CI 1.18-1.35) or metastatic BC (OR 1.59; 95% CI 1.44-1.75) at diagnosis. During a median follow-up of 5.8 years after BC diagnosis 10,900 (14.9%) women died of BC. Risk of BC death was higher among diabetic compared to non-diabetic women (HR 1.36; 95% CI 1.27-1.46). Risk of BC death increased with duration of DM. This supports DM as a risk factor for fatal BC.

PMID: 29318620

DOI: 10.1002/ijc.31250