BACKGROUND
Although trials of post-surgical tamoxifen therapy for patients with breast cancer have convincingly demonstrated reductions in relapse rates and improvements in survival, the optimal duration of therapy is as yet unclear.
PURPOSE
Our objective is to determine whether 2 or 5 years of tamoxifen therapy (20 mg/day orally) is preferable in the treatment of patients with early breast cancer.
METHODS
A randomized trial that was pragmatic in policy, allowing flexibility in primary treatment (i.e., type of surgery) and adjuvant therapy other than tamoxifen (i.e., radiotherapy or chemotherapy), was used to encourage maximum participation of clinicians and patients. This design allowed comparison of the two durations of tamoxifen therapy under conditions in which they would subsequently be applied in clinical practice. The patients were recruited from the oncology departments of participating hospitals in the U.K., Belgium, Poland, and Hong Kong. Those women who had completed an initial 2-year course of tamoxifen therapy after primary surgery and had not experienced a recurrence of breast cancer were asked to consider random assignment either to no further therapy or to an additional 3 years of tamoxifen. Follow-up reports (every 6 months for 3 years after random assignment and then annually) were required for all surviving study subjects. These reports recorded all breast cancer-related events (i.e., locoregional relapse and distant metastasis) or the development of new primary tumors (in the opposite breast or at any other site). For patients who died, the date and cause of death were recorded. Event-free survival; overall survival; and time to locoregional relapse, distant metastasis, or the development of new primary cancers were end points for the analysis. Survival comparisons were made by use of life tables and the logrank test. Reported P values are two-sided.
RESULTS
By December 31, 1994, 2937 patients had accepted random assignment to treatment; 1470 were assigned to stop tamoxifen therapy after having received it for 2 years and 1467 were assigned to continue therapy for an additional 3 years (total, 5 years). An analysis was performed when the target for patient accrual had been reached, although the trial remains open. At a median follow-up of 2 years since the randomization, no difference in survival between the two treatment groups was detected (relative risk = 0.89; 95% confidence interval = 0.69-1.15), but a statistically significant delay in the time to relapse for patients receiving the longer treatment was demonstrated (relative risk = 0.81; 95% confidence interval = 0.69-0.98).
CONCLUSIONS AND IMPLICATIONS
Our results suggest that 5 years may be better than 2 years of tamoxifen therapy, but more evidence regarding the optimal duration of tamoxifen therapy must be obtained.