Lega I C, Fung K, Austin P C, Lipscombe L L
Women's College Research Institute, Women's College Hospital.
Department of Medicine, University of Toronto.
Curr Oncol. 2017 Apr;24(2):e85-e91. doi: 10.3747/co.24.3380. Epub 2017 Apr 27.
The objective of the present study was to use a large, population-based cohort to examine the association between metformin and breast cancer stage at diagnosis while accounting for mammography differences.
We used data from Ontario administrative health databases to identify women 68 years of age or older with diabetes and invasive breast cancer diagnosed from 1 January 2007 to 31 December 2012. Adjusted logistic regression models were used to compare breast cancer stage at diagnosis (stages i and ii vs. iii and iv) between the women exposed and not exposed to metformin. We also examined the association between metformin use and estrogen receptor status, tumour size, and lymph node status in the subset of women for whom those data were available.
We identified 3125 women with diabetes and breast cancer; 1519 (48.6%) had been exposed to metformin before their cancer diagnosis. Median age at breast cancer diagnosis was 76 years (interquartile range: 72-82 years), and mean duration of diabetes was 8.8 ± 5.9 years. In multivariable analyses, metformin exposure was not associated with an earlier stage of breast cancer (odds ratio: 0.98; 95% confidence interval: 0.81 to 1.19). In secondary analyses, metformin exposure was not associated with estrogen receptor-positive breast cancer, tumours larger than 2 cm, or positive lymph nodes.
This population-based study did not show an association between metformin use and breast cancer stage or tumour characteristics at diagnosis. Our study considered older women with long-standing diabetes, and therefore further studies in younger patients could be warranted.
本研究的目的是利用一个基于人群的大型队列，在考虑乳房X线摄影差异的情况下，研究二甲双胍与乳腺癌诊断分期之间的关联。
我们使用安大略省行政卫生数据库的数据，识别出2007年1月1日至2012年12月31日期间诊断出患有糖尿病和浸润性乳腺癌的68岁及以上女性。使用调整后的逻辑回归模型比较暴露于二甲双胍和未暴露于二甲双胍的女性在诊断时的乳腺癌分期（i期和ii期与iii期和iv期）。我们还在可获得这些数据的女性子集中研究了二甲双胍使用与雌激素受体状态、肿瘤大小和淋巴结状态之间的关联。
我们识别出3125名患有糖尿病和乳腺癌的女性；其中1519名（48.6%）在癌症诊断前曾暴露于二甲双胍。乳腺癌诊断时的中位年龄为76岁（四分位间距：72 - 82岁），糖尿病平均病程为8.8±5.9年。在多变量分析中，二甲双胍暴露与乳腺癌的早期分期无关（比值比：0.98；95%置信区间：0.81至1.19）。在二次分析中，二甲双胍暴露与雌激素受体阳性乳腺癌、大于2 cm的肿瘤或阳性淋巴结无关。
这项基于人群的研究未显示二甲双胍使用与乳腺癌诊断分期或肿瘤特征之间存在关联。我们的研究纳入了患有长期糖尿病的老年女性，因此可能有必要对年轻患者进行进一步研究。
