Oral Contraceptives and the Risk of Breast Cancer


Synopsis: Old, no longer used, high-dose oral contraceptives may increase the risk of breast cancer in first-degree relatives of women who had breast cancer.

Source: Grabrick DM, et al. JAMA 2000;284:1791-1798.

Grabrick and associates from the mayo clinic conducted a follow-up study of the association between oral contraceptive (OC) use and the risk of breast cancer in a cohort of women with a family history of breast cancer. This is a study of families using a cohort first identified between 1944 and 1952 (The Minnesota Breast Cancer Family Study), studied again from 1991-1996, and now in this most recent follow-up, these families were reassessed by telephone interviews. From the original 544 families, a total of 426 families could be updated, yielding a total of 6150 women. In the 6150 women, 153 cases of breast cancer occurred in blood relatives and 86 in marry-ins. In the entire cohort, the relative risk of breast cancer associated with use of oral contraceptives was 1.4 (confidence interval [CI] = 1.0-2.0), not achieving statistical significance. Grabrick et al assessed the effect of inheritance by assessing the risk by relationships to the women initially identified with breast cancer from 1944 to 1952. Sisters and daughters who had ever used OCs had a relative risk of 3.3 (CI = 1.6-6.7) compared with sisters and daughters who were never users. This was not seen with granddaughters, nieces, or marry-ins. No differences were noted for duration of use or age at first use. Among 132 high-risk families (³ 3 breast or ovarian cancers among blood relatives), even greater relative risks were observed, although the CI was wide because of small numbers. The increased risks were present only in relatives who used OCs prior to 1975, a period during which OCs contained high doses of estrogen (> 50 mg ethinyl estradiol) and progestin. These results remained after adjustments for breast cancer risk factors and for number of mammograms.

Comment by Sarah L. Berga, MD

This is yet another study attempting to understand who really is at risk for breast cancer if they use exogenous steroids. In this scenario, the exogenous hormones are OCs and the subjects are relatives of women who had breast cancer roughly 50 years ago. Although Grabrick et al chose an interesting approach, the data are nonetheless inconclusive. How, then, does one put this study into perspective? First, the study looked at the incidence of breast cancer, not death from breast cancer. The choice of this end point, however pragmatic, obviously limits the usefulness of the study. Second, Grabrick et al did not ascertain the doses in the OCs and, therefore, used year of use as a surrogate for dose. The customary dose prior to 1975 was 50 µg or more of ethinyl estradiol. No one prescribes this dose of OCs any longer, mostly because of the side effects and the higher risk of venous thromboembolic events. Thus, it is too soon to know if there is any excess risk of breast cancer in women with a family history of breast cancer if they use 20 µg oral contraceptives. Oddly, there was no increase in risk with extended duration of use. This may reflect the difficulty of finding a sample size large enough to allow such relationships to be addressed. There did seem to be a dose relationship, however, between genetic risk and ever use, suggesting that women with a very strong family history of breast cancer appear to more susceptible to agents that are safe to use in everyone else.

Most physicians would not find the above information overly alarming because most women do not have a strong family history of breast cancer. However, most patients would be upset, and many might be tempted to stop using OCs as their method of birth control. Further, many women harbor fears that they are at excess risk of getting breast cancer, even if they do not have a strong family history. Reassurance sometimes helps in this context. But assuming that there is a kernel of truth in the above data set, the question facing the practicing physician is what to do for the women who truly does have a strong family history, especially the woman who has tested positive for BRCA1 or BRCA2. Should these women be counseled to not use oral contraceptives? Should they be counseled to do something else such as take tamoxifen premenopausally? Should they be offered a medical or surgical oophorectomy, prophylactic bilateral mastectomy, or both? In an era of managed cost, who decides? There are no good answers to these questions. The sad fact is that we have no clue as to what to recommend. All of the choices seem bad. Personally, I do not find the present study of sufficient strength that I would deny the use of very low-dose OCs to women with a family history of breast cancer. But I would also find it difficult to advocate for their use if the family history is compelling.

Comment by Leon Speroff, MD

Grabrick et al state in their discussion that "our results suggest that the use of OCs in women with a strong family history of breast cancer may further elevate their breast cancer risk." However, what this study indicates is that high-dose OCs, no longer used, may increase the risk of breast cancer in women with family histories of breast cancer. To be sure, the numbers of users after 1975 were not large and the women were younger, having not experienced yet their entire life-long risk of breast cancer. Nevertheless, the only positive result is limited to high-dose OCs.

I am especially disturbed by the discussion that refers to the previous study concluding that OC use reduces the risk of ovarian cancer in women with BRCA1 or BRCA2 mutations and argues that the current results indicate that these same women would be at higher risk of breast cancer with the use of OCs.1 They further support this warning by referring to a study that suggested that OCs increase the risk of breast cancer in carriers of the BRCA mutations.2 The BRCA ovarian cancer study was performed in women using current low-dose OCs; the current study found an increased risk only with old high-dose pills, not with current low-dose pills. The BRCA breast cancer study involved very small numbers and did not achieve statistical significance. Grabrick et al do not mention these important qualifications. This is yet another example of careless, selective reporting. Therefore, the advice that women with BRCA mutations should avoid low-dose OCs is unfounded and inappropriate, potentially depriving these women of an important benefit.

The conclusions of this study are derived from small numbers. The increased relative risk of 3.3 in sisters and daughters was based on a total of 13 ever users and 25 never users. The accompanying editorial3 states that the data offer "strong support" for an increased effect in the presence of genetic risk, based upon an increasing relative risk when families had three to five or more members with breast or ovarian cancer. However, these conclusions were based upon 10 users and 16 never users (³ 3 cancers) and six users and three never users (³ 5 cancers). Although the conclusions were statistically significant, the confidence intervals were wide, 1.6-6.7 and 2.3-56.4, respectively. Thus, the conclusions are imprecise due to small numbers, hardly robust support as implied by the editorial.

Therefore, I am not convinced that this report is "bad news" for women with a family history of breast cancer, as the editorial suggests. The collaborative re-analysis of the world’s data concluded that OC use does not further increase the risk of breast cancer with positive family histories of breast cancer.4 I believe that low-dose OCs should continue to be offered as an appropriate choice for women with family histories of breast and ovarian cancers.


1. Narod SA, et al. N Engl J Med 1998;339:424-428.

2. Ursin G, et al. Cancer Res 1997;57:3678-3681.

3. Burke W. JAMA 2000;284:1837-1838.

4. Collaborative Group on Hormonal Factors in Breast Cancer. Lancet 1996;347:1713-1727.