Tamoxifen's effect in women with breast cancer

Abstract

Steven Narod and colleagues' study (Dec 2, p 1876)1Narod S A Brunet J-S Ghadirian P et al.Tamoxifen and the risk of contralateral breast cancer in BRCA1 and BRCA2 mutation carriers: a case-control study.Lancet. 2000; 356: 1876-1881Summary Full Text Full Text PDF PubMed Scopus (503) Google Scholar can be seen as a useful initial investigation into the potential protective effect of tamoxifen in women with BRCA1 and BRCA2 mutations. They note that the odds of previous tamoxifen use were 50% lower in women with contralateral breast cancer than in those with unilateral cancer. However we have some concerns about the collection of data and study population.Questionnaires were used as the sole method of data collection. These were carried out in different ways, and by different people, with some centres using telephone interviews; leading to a potential lack of consistency between centres. The questionnaires were completed an average of 11·8 years after initial diagnosis, and recall bias can affect recall of details of drug use. We would be interested to know why medical records were not used as a more accurate source of information.As explained by Narod and colleagues, only prevalent cases were studied, since genetic screening was necessary. This approach excludes women who had died as a result of familial breast cancer. Contralateral breast cancer has a higher mortality rate than unilateral cancer,2Leis HP Selective, elective, prophylactic contralateral mastectomy.Cancer. 1971; 28: 956-961Crossref PubMed Scopus (39) Google Scholar and, therefore, more cases than controls were probably excluded because of premature death. The investigators' main conclusions are based on the finding that more controls than cases used tamoxifen after initial diagnosis. The women who died— potential cases and controls—are likely to have used tamoxifen, but in what proportions it is impossible to know, as is the proportion of cases and controls who died. Had the women who died been included, the findings might have differed sub-stantially and diluted the observed effect. Tamoxifen might, therefore, have no protective effect in mutation carriers, and could even be harmful. If so, potential cases might have died, which would lower the number of women reporting tamoxifen use.If genetic analysis were used routinely at initial diagnosis of familial breast cancer, incident cases could be included in prospective case-control studies. Women who died during the study would be included and findings more reliable. Steven Narod and colleagues' study (Dec 2, p 1876)1Narod S A Brunet J-S Ghadirian P et al.Tamoxifen and the risk of contralateral breast cancer in BRCA1 and BRCA2 mutation carriers: a case-control study.Lancet. 2000; 356: 1876-1881Summary Full Text Full Text PDF PubMed Scopus (503) Google Scholar can be seen as a useful initial investigation into the potential protective effect of tamoxifen in women with BRCA1 and BRCA2 mutations. They note that the odds of previous tamoxifen use were 50% lower in women with contralateral breast cancer than in those with unilateral cancer. However we have some concerns about the collection of data and study population. Questionnaires were used as the sole method of data collection. These were carried out in different ways, and by different people, with some centres using telephone interviews; leading to a potential lack of consistency between centres. The questionnaires were completed an average of 11·8 years after initial diagnosis, and recall bias can affect recall of details of drug use. We would be interested to know why medical records were not used as a more accurate source of information. As explained by Narod and colleagues, only prevalent cases were studied, since genetic screening was necessary. This approach excludes women who had died as a result of familial breast cancer. Contralateral breast cancer has a higher mortality rate than unilateral cancer,2Leis HP Selective, elective, prophylactic contralateral mastectomy.Cancer. 1971; 28: 956-961Crossref PubMed Scopus (39) Google Scholar and, therefore, more cases than controls were probably excluded because of premature death. The investigators' main conclusions are based on the finding that more controls than cases used tamoxifen after initial diagnosis. The women who died— potential cases and controls—are likely to have used tamoxifen, but in what proportions it is impossible to know, as is the proportion of cases and controls who died. Had the women who died been included, the findings might have differed sub-stantially and diluted the observed effect. Tamoxifen might, therefore, have no protective effect in mutation carriers, and could even be harmful. If so, potential cases might have died, which would lower the number of women reporting tamoxifen use. If genetic analysis were used routinely at initial diagnosis of familial breast cancer, incident cases could be included in prospective case-control studies. Women who died during the study would be included and findings more reliable. Tamoxifen's effect in women with breast cancerAuthors' reply Full-Text PDF