Correspondence
Oral Contraceptives and Hereditary Ovarian Cancer
N Engl J Med 1999; 340:59-60January 7, 1999
- Article
To the Editor:
Narod and colleagues (Aug. 13 issue)1 suggest that oral-contraceptive therapy reduces the risk of ovarian cancer among women with mutations in the BRCA1 or BRCA2 gene. However, some possible biases should be considered.
First, the characteristics of the 207 women with ovarian cancer and the 161 control women were not similar, in that all of the patients had a BRCA1 or BRCA2 mutation but only 53 of the 95 women in the control group who were tested were carriers of the same mutations as their sisters. Even if the 66 control women who were not tested for mutations were all carriers of a BRCA1 or BRCA2 mutation, the frequency of carriers in the control group would be 74 percent. Therefore, the most important risk factor for ovarian cancer was not equally prevalent among the patients and the controls. Such selection bias could lead to an overestimation of the magnitude of the protective effect attributed to oral-contraceptive therapy.
Second, since the patients and the controls were not paired (even though all the controls were sisters of the patients), the fact that there were zero to seven controls for each patient could lead to selection bias. For example, if patients with a family history of ovarian or breast cancer had no sisters or only one sister in the study and the patients with no such history had five or more sisters in the study, the prevalence of a family history could differ significantly between the two groups, therefore also leading to an overestimation of the protective effect attributed to oral-contraceptive therapy.
1 ReferencesRafael Marques de Souza
Rua Heller 61/81, CEP 93510-330 Novo Hamburgo, RS, BrazilAnderson Rech Lazzaron
Universidade Federal do Rio Grande do Sul, CEP 90046-900 Porto Alegre, RS, Brazil1
Narod SA ,Risch H ,Moslehi R , et al. Oral contraceptives and the risk of hereditary ovarian cancer. N Engl J Med 1998;339:424-428
Full Text | Web of Science | Medline
To the Editor:
Although the case–control study by Narod et al. suggests a substantial protective effect of oral contraceptives against hereditary ovarian cancer, a bias may have arisen from the fact that more patients than control women had a history of breast cancer (30 percent vs. 18 percent). Given that the epidemiologic relation between contraceptive hormones and breast cancer has received widespread attention,1 some women with a history of early breast cancer may have perceived that oral contraceptives would unfavorably affect their prognosis or risk of recurrence and avoided using them. The finding that women with a history of breast cancer and those without such a history had a similar mean duration of oral-contraceptive use does not fully reassure one of the absence of this bias, because these measures of duration were not standardized for age or other correlates of breast cancer. What are the estimated relative risks if women with a history of breast cancer are excluded?
1 ReferencesRobert C. Kaplan, M.S., Ph.C.
University of Washington, Seattle, WA 98195-72361
The Collaborative Group on Hormonal Factors in Breast Cancer
CrossRef | Web of Science | Medline
Author/Editor Response
The authors reply:
To the Editor: De Souza and Lazzaron point out that more women with ovarian cancer than control women were carriers of mutations. This is not a result but a direct consequence of our study design, whereby we chose to include noncarrier sisters as controls. The reasons for this design are explained in detail in the article. The inclusion of noncarrier control women should attenuate the relative risk associated with oral-contraceptive therapy, rather than accentuating it, as de Souza and Lazzaron assert. When we restricted our control group to women who were carriers, the odds ratio declined from 0.4 to 0.3.
Kaplan suggests that women with breast cancer may have avoided oral-contraceptive use and should have been excluded. The women with breast cancer had taken oral contraceptives for an average of 2.9 years, as compared with 3.2 years for women without breast cancer. Among the women with breast cancer, 97 percent of oral-contraceptive use occurred before the diagnosis of cancer. Carriers of BRCA1 or BRCA2 mutations have a very high lifetime risk of breast cancer. If the use of oral contraceptives increases their risk of breast cancer, then the exclusion of these women will leave a subgroup with a lower-than-expected use of oral contraceptives. In any case, in our data set the odds ratio for ovarian cancer associated with any use of oral contraceptives, after the exclusion of both patients and controls with breast cancer, was 0.5. This result is virtually the same as when the women with breast cancer were included.
Steven A. Narod, M.D.
Jean-Sébastien Brunet, M.Sc.
Women's College Hospital, Toronto, ON M5G 1N8, Canadafor the Hereditary Ovarian Cancer Clinical Study Group
