By Robert L. Coleman, MD
Professor, University of Texas; M.D. Anderson Cancer Center, Houston
Dr. Coleman reports no financial relationships relevant to this field of study.
This article originally appeared in the March 2014 issue of OB/GYN Clinical Alert.
SYNOPSIS:The association between oral contraceptive use and ovarian or breast cancer in BRCA1 or BRCA2 mutation carriers are qualitatively similar to associations reported in the general population. Oral contraceptive pill use is inversely associated with ovarian cancer risk. However, it is also associated with a modest, but not statistically significant, increased risk for breast cancer. The analysis was unable to provide conclusive recommendations as to their use as preventive measures given these and other unmeasured risks. However, oral contraceptive pills appear safe for contraception in this population.
SOURCE:Moorman PG, et al. Oral contraceptives and risk of ovarian cancer and breast cancer among high-risk women: A systematic review and meta-analysis. J Clin Oncol 2013;31:4188-4198.
Risks for ovarian and breast cancer are substantially elevated in women who carry germline mutations in BRCA1/2. The most effective method of cancer prevention is surgical resection; however, in those who wish to preserve fertility options, non-permanent prevention strategies are desired. Currently, non-invasive screening is unproven even in this high-risk cohort. Oral contraception pills (OCP) have been documented to reduce ovarian cancer risk in the general population and the magnitude of effect is related to the duration of use. The current study was conducted to analyze the known datasets of OCP use in high-risk women (i.e., carriers of BRCA1/2 or with a strong family history) for ovarian and breast cancer risk. The meta-analysis considered 6476 unique citations examining ovarian and breast cancer risk and settled on six addressing ovarian cancer risk and eight addressing breast cancer risk. Among germline mutation carriers combined, the meta-analysis demonstrated an inverse association between OCP use and ovarian cancer (odds ratio [OR], 0.58; 95% confidence interval [CI], 0.46-0.73) and a non-statistically significant association with breast cancer (OR, 1.21; 95% CI, 0.93-1.58). Findings were similar when examining BRCA1 and BRCA2 mutation carriers separately. The data were inadequate to perform a meta-analyses examining duration or timing of use. Additionally, there were four studies examining risk for ovarian cancer and three for breast cancer among women with a family history of ovarian or breast cancer. However, differences between studies precluded combining the data for meta-analyses, and no overall pattern could be discerned. The authors concluded that ever use of OCPs in women carrying a germline mutation in BRCA1 or BRCA2 was similar to that demonstrated in studies of population-risk patients. However, risk/benefit could not be directly addressed precluding a recommendation for their use for prevention of ovarian cancer.
COMMENTARY
"Oral contraceptive use had no significant effect on ovarian, breast cancer risk in BRCA1/2 carries" was the headline in a recent medical periodical highlighting this specific article. As can be appreciated, the sound bite is misleading, and although it gets one aspect correct (impact on breast cancer), it is stated with the intent to highlight the lack of protection by OCPs for ovarian cancer (false) and breast cancer (false) development in this high-risk group. While the authors clearly demonstrate an inverse protective effect of OCP use and ovarian cancer, the concern was not in protection of breast cancer but rather that OCP use would increase breast cancer risk, particularly in this patient cohort of individuals at substantially higher risk of breast cancer. Previous reports have raised concerns that OCP use may increase breast cancer risk in the general population.1,2 The authors demonstrated an OR for breast cancer risk of > 1.0, but it was not statistically significant. Overall, this should be reassuring, yet, it was concluded that there was a "non-statistically significant" association of OCP use and the development of breast cancer.
This experience raises two take-home messages: first, headlines and sound bites may be very misleading and should be reproduced with caution; and second, data from observational studies and meta-analyses are hypothesis-generating and should be limited in their scope to these activities. Although, randomized controlled trials are the gold standard in assessing effect, such studies involving an intervention like OCPs are impractical. However, properly designed and monitored cohort studies, in this setting, can provide strong estimates of effect. Third, meta-analyses are tricky to perform properly, and heterogeneity in study design, patient cohorts, treatment or intervention used, follow-up, and confounders of the individual trials included in the exercise make it extremely difficult to assess questions of risk.3 As was appreciated in this current study, the ratio of assessable trials filling the eligibility criteria was about 1:1000 in the reported literature. Finally, we are cautioned to accept even the author's conclusion based on the presented data. In this manuscript, the authors state "There is insufficient evidence to recommend oral contraceptive use as a chemoprevention strategy in high-risk women, if they otherwise would not be taking them for contraception." At face value this may be true, but the intent was not to diminish the effect seen in their use, but rather to provide a statement regarding the risks of OCP use (which they could not assess due to trial heterogeneity in their sample set) and benefits (which they did evaluate in their sample set).4 As is often the case, medical practice is governed by imperfect data and we are left to critically interpret the information before us; this should be done with caution.
REFERENCES
- Kay CR, Hannaford PC. Br J Cancer 1988;58:675-680.
- Lee NC, et al. J Natl Cancer Inst 1987;79:1247-1254.
- Ioannidis JP, Lau J. Jt Comm J Qual Improv 1999;25:462-469.
- Ness RB, et al. Epidemiol 2001;12:307-312.
