Year: 2012
Type: article
Abstract: For the sake of brevity, we cited comprehensive reviews in our article ( 1 ), and therefore, we regret that we failed to specifically cite a previously published observational study by Abraham et al. ( 2 ), which was also authored by Drs Pharoah and Caldas, and, similar to the Arimidex, Tamoxifen, Alone or in Combination (ATAC) and Breast International Group (BIG) 1-98 studies ( 1 , 3 ), failed to detect an association between CYP2D6 genotype and response to tamoxifen. Pharoah et al. appear to agree with our conclusion that CYP2D6 genotype does not influence tamoxifen activity in women with estrogen receptor–positive breast cancer, and that CYP2D6 genetic testing is not indicated for such women. Pharoah et al. question the need for use of specimens from a randomized trial to test the clinical utility of a biomarker. We recognize the value of data from well-controlled nonrandomized observational studies. However, we maintain, as proposed by Simon et al. ( 4 ), that biomarker data generated from prospective, and preferably, randomized trials are likely to be less biased than those from simple observational studies. In the latter, the cohort, endpoint, and biomarker comparisons are often not specified before the analysis is undertaken. A nontreated control group, or a control group treated with aromatase inhibitors (which are not affected by CYP2D6 genotypes), which was the case in ATAC and BIG 1-98 trials, provides an opportunity to validate that any apparent positive findings are unique to tamoxifen. Moreover, the quality of prospectively collected clinical outcomes data from trials, leading to drug registration, reduce unrecognized biases that may exist in observational studies, in which the data are not as accurately recorded.
DOI: 10.1093/jnci/djs302
Authors: J. M. Rae, D. F. Hayes, J. Cuzick, I. Sestak, M. Dowsett
Language: English
Subfield: Pharmacology
Domain: Life Sciences
Open Access status: bronze
OpenAlex ID: W4245134682