Anastrozole Is Superior to Tamoxifen as First-Line Therapy in Hormone
Receptor-Positive Advanced Breast Carcinoma I was surprised to read the recent article by Bonneterre et al. 1 The data presented in that article are the result of a retrospective analysis of a selected number of patients from two already published clinical trials that had previously reported conflicting results. The North American trial 2 had reported a significant advantage for anastrozole over tamoxifen in time to progression (median TTP: 11.1 vs. 5.6 months, respectively), whereas the larger European trial 3 did not show a significant improvement (median TTP: 8.2 months for anastrozole vs. 8.3 months for tamoxifen). This discrepancy was rationalized based on the fact that the North American trial enrollment was about 85% hormone receptor-positive patients, compared with about 44% in the European trial. However, statistical significance appears not to have been reached either in this subgroup in the European trial (TTP: 8.9 months for anastrozole vs. 7.8 months for tamoxifen). 3 Of further concern is the trend toward a worse TTP in the European trial in patients with receptor-unknown tumors (7.3 months with anastrozole compared with 8.3 months with tamoxifen).
In the European trial, P values were not provided for the subgroup analyses, and we can only infer from the data. However, notwithstanding that any retrospective analysis, even from a prospective trial, does not create Level I evidence, but at best Level III evidence, and as such weakens the conclusions that one can draw from the data, 4 it was resolved to combine the data from both trials to give more power to the analysis. The need for such a strategy appears questionable since neither trial was designed as a superiority trial in the first place, but as an equivalence trial. As prospectively planned, both trials achieved their primary objective, namely to show equivalence of tamoxifen and anastrozole. Although, statistically speaking, this is the only allowable conclusion, both the paper by Bonneterre et al. 1 and the original publications attempt to show superiority, and the latter have been criticized for this. 5,6 In the retrospective analysis, the conclusions of superiority that are reached are inconsistent with those from the original prospective trials, and they do not explain why the North American trial succeeded in achieving statistically significant superiority in hormone receptor-positive patients whereas the European trial did not although both trials enrolled a similar number of such patients (145 vs. 156 for anastrozole and tamoxifen, respectively, in the former, and 153 vs. 144 for anastrozole and tamoxifen, respectively, in the latter). An additional concern about the claim that anastrozole is superior to tamoxifen is the lack of survival data in the article, although the two-year survival data for both studies are available on the Food and Drug Administration web site. 7 In fact, the Center for Drug Evaluation and Research Medical Review reports less 2442